nimbus-eth1/nimbus/vm/interpreter/opcodes_impl.nim

1093 lines
34 KiB
Nim
Raw Normal View History

Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
# Nimbus
# Copyright (c) 2018-2019 Status Research & Development GmbH
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
# 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.
import
strformat, times, sets, sequtils, options,
chronicles, stint, nimcrypto, stew/ranges/ptr_arith, eth/common,
./utils/[macros_procs_opcodes, utils_numeric],
./gas_meter, ./gas_costs, ./opcode_values,
Transaction: EVMC fix, `CREATE2` salt is a 256-bit blob not a number This changes fixes a bug in `CREATE2` ops when used with EVMC. Because it changes the salt type, it affects non-EVMC code as well. The salt was passed through EVMC with the wrong byte order, although this went unnoticed as the Nimbus host flipped the byte order before using it. This was found when running Nimbus with third-party EVM, ["evmone"](https://github.com/ethereum/evmone). There are different ways to remedy this. If treated as a number, Nimbus EVM would byte-flip the value when calling EVMC, then Nimbus host would flip the received value. Finally, it would be flipped a third time when generating the address in `generateSafeAddress`. The first two flips can be eliminated by negotiation (like other numbers), but there would always be one flip. As a bit pattern, Nimbus EVM would flip the same way it does when dealing with hashes on the stack (e.g. with `getBlockHash`). Nimbus host wouldn't flip at all - and when using third-party EVMs there would be no flips in Nimbus. Because this value is not for arithmetic, any bit pattern is valid, and there shouldn't be any flips when using a third-party EVM, the bit-pattern interpretation is favoured. The only flip is done in Nimbus EVM (and might be eliminated in an optimised version). As suggested, we'll define a new "opaque 256 bits" type to hold this value. (Similar to `Hash256`, but the salt isn't necessarily a hash.) Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-06-09 22:32:42 +00:00
".."/[memory, stack, code_stream, computation, state, types],
".."/../[errors, constants, forks, utils],
../../db/[db_chain, accounts_cache]
when defined(evmc_enabled):
import ../evmc_api, ../evmc_helpers, evmc/evmc
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
logScope:
topics = "opcode impl"
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
# ##################################
# Syntactic sugar
template accessGas(c: Computation, address: EthAddress): GasInt =
c.accessGas(ColdAccountAccessCost, WarmStorageReadCost, address)
2021-01-13 11:17:38 +00:00
template accessGas(c: Computation, coldAccessGas, warmAccessGas: GasInt): GasInt =
# Simulate making the `address` argument default `= c.msg.contractAddress`.
# Works around Nim 1.2.10 compiler's "internal error: environment misses: c".
c.accessGas(coldAccessGas, warmAccessGas, c.msg.contractAddress)
template accessGas(c: Computation, coldAccessGas, warmAccessGas: GasInt,
address: EthAddress): GasInt =
c.accessGas(address, coldAccessGas, warmAccessGas)
proc accessGas(c: Computation, address = c.msg.contractAddress,
coldAccessGas, warmAccessGas: GasInt): GasInt {.inline.} =
when evmc_enabled:
if c.host.accessAccount(address) == EVMC_ACCESS_COLD:
coldAccessGas
else:
warmAccessGas
else:
c.vmState.mutateStateDB:
if not db.inAccessList(address):
db.accessList(address)
return coldAccessGas
else:
return warmAccessGas
proc accessGas(c: Computation, coldAccessGas, warmAccessGas: GasInt,
2022-04-08 04:54:11 +00:00
slot: UInt256): GasInt {.inline.} =
when evmc_enabled:
if c.host.accessStorage(c.msg.contractAddress, slot) == EVMC_ACCESS_COLD:
coldAccessGas
else:
warmAccessGas
else:
c.vmState.mutateStateDB:
if not db.inAccessList(c.msg.contractAddress, slot):
db.accessList(c.msg.contractAddress, slot)
return coldAccessGas
else:
return warmAccessGas
2020-12-11 10:51:17 +00:00
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
template push(x: typed) {.dirty.} =
## Push an expression on the computation stack
c.stack.push x
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
# ##################################
# 0s: Stop and Arithmetic Operations
op add, inline = true, lhs, rhs:
## 0x01, Addition
push: lhs + rhs
op mul, inline = true, lhs, rhs:
## 0x02, Multiplication
push: lhs * rhs
op sub, inline = true, lhs, rhs:
## 0x03, Substraction
push: lhs - rhs
op divide, inline = true, lhs, rhs:
## 0x04, Division
push:
2022-04-08 04:54:11 +00:00
if rhs == 0: zero(UInt256) # EVM special casing of div by 0
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
else: lhs div rhs
op sdiv, inline = true, lhs, rhs:
## 0x05, Signed division
var r: UInt256
if rhs != 0:
var a = lhs
var b = rhs
var signA, signB: bool
extractSign(a, signA)
extractSign(b, signB)
r = a div b
setSign(r, signA xor signB)
push(r)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op modulo, inline = true, lhs, rhs:
## 0x06, Modulo
push:
2022-04-08 04:54:11 +00:00
if rhs == 0: zero(UInt256)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
else: lhs mod rhs
op smod, inline = true, lhs, rhs:
## 0x07, Signed modulo
var r: UInt256
if rhs != 0:
var sign: bool
var v = lhs
var m = rhs
extractSign(m, sign)
extractSign(v, sign)
r = v mod m
setSign(r, sign)
push(r)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op addmod, inline = true, lhs, rhs, modulus:
## 0x08, Modulo addition
## Intermediate computations do not roll over at 2^256
push:
if modulus == 0: zero(UInt256) # EVM special casing of div by 0
else: addmod(lhs, rhs, modulus)
op mulmod, inline = true, lhs, rhs, modulus:
## 0x09, Modulo multiplication
## Intermediate computations do not roll over at 2^256
push:
if modulus == 0: zero(UInt256) # EVM special casing of div by 0
else: mulmod(lhs, rhs, modulus)
op exp, inline = true, base, exponent:
## 0x0A, Exponentiation
c.gasMeter.consumeGas(
c.gasCosts[Exp].d_handler(exponent),
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
reason="EXP: exponent bytes"
)
push:
if base.isZero:
if exponent.isZero:
# https://github.com/ethereum/yellowpaper/issues/257
# https://github.com/ethereum/tests/pull/460
# https://github.com/ewasm/evm2wasm/issues/137
1.u256
else:
zero(UInt256)
else:
base.pow(exponent)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op signExtend, inline = false, bits, value:
## 0x0B, Sign extend
## Extend length of twos complement signed integer.
var res: UInt256
if bits <= 31.u256:
let
2019-02-23 14:42:23 +00:00
one = 1.u256
2019-11-13 14:49:39 +00:00
testBit = bits.truncate(int) * 8 + 7
2019-02-23 14:42:23 +00:00
bitPos = one shl testBit
mask = bitPos - one
if not isZero(value and bitPos):
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
res = value or (not mask)
else:
res = value and mask
else:
res = value
push: res
# ##########################################
# 10s: Comparison & Bitwise Logic Operations
op lt, inline = true, lhs, rhs:
## 0x10, Less-than comparison
push: (lhs < rhs).uint.u256
op gt, inline = true, lhs, rhs:
## 0x11, Greater-than comparison
push: (lhs > rhs).uint.u256
op slt, inline = true, lhs, rhs:
## 0x12, Signed less-than comparison
push: (cast[Int256](lhs) < cast[Int256](rhs)).uint.u256
op sgt, inline = true, lhs, rhs:
## 0x13, Signed greater-than comparison
push: (cast[Int256](lhs) > cast[Int256](rhs)).uint.u256
op eq, inline = true, lhs, rhs:
## 0x14, Signed greater-than comparison
push: (lhs == rhs).uint.u256
op isZero, inline = true, value:
## 0x15, Check if zero
push: value.isZero.uint.u256
op andOp, inline = true, lhs, rhs:
## 0x16, Bitwise AND
push: lhs and rhs
op orOp, inline = true, lhs, rhs:
## 0x17, Bitwise AND
push: lhs or rhs
op xorOp, inline = true, lhs, rhs:
## 0x18, Bitwise AND
push: lhs xor rhs
op notOp, inline = true, value:
## 0x19, Check if zero
push: value.not
op byteOp, inline = true, position, value:
## 0x20, Retrieve single byte from word.
2019-11-13 14:49:39 +00:00
let pos = position.truncate(int)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
push:
2022-04-08 04:54:11 +00:00
if pos >= 32 or pos < 0: zero(UInt256)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
else:
when system.cpuEndian == bigEndian:
cast[array[32, byte]](value)[pos].u256
else:
cast[array[32, byte]](value)[31 - pos].u256
# ##########################################
# 20s: SHA3
op sha3, inline = true, startPos, length:
## 0x20, Compute Keccak-256 hash.
2019-05-01 12:21:46 +00:00
let (pos, len) = (startPos.safeInt, length.safeInt)
2019-04-25 15:33:26 +00:00
2018-09-18 19:57:19 +00:00
if pos < 0 or len < 0 or pos > 2147483648:
raise newException(OutOfBoundsRead, "Out of bounds memory access")
c.gasMeter.consumeGas(
c.gasCosts[Op.Sha3].m_handler(c.memory.len, pos, len),
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
reason="SHA3: word gas cost"
)
c.memory.extend(pos, len)
let endRange = min(pos + len, c.memory.len) - 1
if endRange == -1 or pos >= c.memory.len:
push(EMPTY_SHA3)
else:
push:
keccak256.digest c.memory.bytes.toOpenArray(pos, endRange)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
# ##########################################
# 30s: Environmental Information
proc writePaddedResult(mem: var Memory,
2022-04-08 04:54:11 +00:00
data: openArray[byte],
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
memPos, dataPos, len: Natural,
paddingValue = 0.byte) =
2019-03-18 10:24:25 +00:00
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
mem.extend(memPos, len)
2018-08-09 16:15:39 +00:00
let dataEndPosition = dataPos.int64 + len - 1
2018-09-18 00:35:41 +00:00
let sourceBytes = data[min(dataPos, data.len) .. min(data.len - 1, dataEndPosition)]
mem.write(memPos, sourceBytes)
2019-03-18 10:24:25 +00:00
2019-03-16 14:42:06 +00:00
# Don't duplicate zero-padding of mem.extend
2019-03-19 09:43:38 +00:00
let paddingOffset = min(memPos + sourceBytes.len, mem.len)
let numPaddingBytes = min(mem.len - paddingOffset, len - sourceBytes.len)
if numPaddingBytes > 0:
# TODO: avoid unnecessary memory allocation
mem.write(paddingOffset, repeat(paddingValue, numPaddingBytes))
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op address, inline = true:
## 0x30, Get address of currently executing account.
push: c.msg.contractAddress
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op balance, inline = true:
## 0x31, Get balance of the given account.
let address = c.stack.popAddress()
2020-01-16 15:07:04 +00:00
push: c.getBalance(address)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op origin, inline = true:
## 0x32, Get execution origination address.
2020-01-16 10:23:51 +00:00
push: c.getOrigin()
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op caller, inline = true:
## 0x33, Get caller address.
push: c.msg.sender
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op callValue, inline = true:
## 0x34, Get deposited value by the instruction/transaction
## responsible for this execution
push: c.msg.value
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op callDataLoad, inline = false, startPos:
## 0x35, Get input data of current environment
2019-08-20 09:26:27 +00:00
let start = startPos.cleanMemRef
if start >= c.msg.data.len:
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
push: 0
return
2019-08-20 09:26:27 +00:00
# If the data does not take 32 bytes, pad with zeros
let endRange = min(c.msg.data.len - 1, start + 31)
2019-08-20 09:26:27 +00:00
let presentBytes = endRange - start
# We rely on value being initialized with 0 by default
var value: array[32, byte]
value[0 .. presentBytes] = c.msg.data.toOpenArray(start, endRange)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2019-08-20 09:26:27 +00:00
push: value
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op callDataSize, inline = true:
## 0x36, Get size of input data in current environment.
push: c.msg.data.len.u256
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op callDataCopy, inline = false, memStartPos, copyStartPos, size:
## 0x37, Copy input data in current environment to memory.
# TODO tests: https://github.com/status-im/nimbus/issues/67
let (memPos, copyPos, len) = (memStartPos.cleanMemRef, copyStartPos.cleanMemRef, size.cleanMemRef)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
c.gasMeter.consumeGas(
c.gasCosts[CallDataCopy].m_handler(c.memory.len, memPos, len),
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
reason="CallDataCopy fee")
c.memory.writePaddedResult(c.msg.data, memPos, copyPos, len)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2019-03-16 14:42:06 +00:00
op codeSize, inline = true:
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
## 0x38, Get size of code running in current environment.
push: c.code.len
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2019-03-16 14:42:06 +00:00
op codeCopy, inline = false, memStartPos, copyStartPos, size:
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
## 0x39, Copy code running in current environment to memory.
# TODO tests: https://github.com/status-im/nimbus/issues/67
let (memPos, copyPos, len) = (memStartPos.cleanMemRef, copyStartPos.cleanMemRef, size.cleanMemRef)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
c.gasMeter.consumeGas(
c.gasCosts[CodeCopy].m_handler(c.memory.len, memPos, len),
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
reason="CodeCopy fee")
c.memory.writePaddedResult(c.code.bytes, memPos, copyPos, len)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op gasprice, inline = true:
## 0x3A, Get price of gas in current environment.
2020-01-16 10:23:51 +00:00
push: c.getGasPrice()
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op extCodeSize, inline = true:
## 0x3b, Get size of an account's code
2020-01-16 15:15:20 +00:00
let address = c.stack.popAddress()
push: c.getCodeSize(address)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op extCodeCopy, inline = true:
## 0x3c, Copy an account's code to memory.
let address = c.stack.popAddress()
let (memStartPos, codeStartPos, size) = c.stack.popInt(3)
let (memPos, codePos, len) = (memStartPos.cleanMemRef, codeStartPos.cleanMemRef, size.cleanMemRef)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
c.gasMeter.consumeGas(
c.gasCosts[ExtCodeCopy].m_handler(c.memory.len, memPos, len),
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
reason="ExtCodeCopy fee")
let codeBytes = c.getCode(address)
c.memory.writePaddedResult(codeBytes, memPos, codePos, len)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op returnDataSize, inline = true:
## 0x3d, Get size of output data from the previous call from the current environment.
push: c.returnData.len
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op returnDataCopy, inline = false, memStartPos, copyStartPos, size:
## 0x3e, Copy output data from the previous call to memory.
let (memPos, copyPos, len) = (memStartPos.cleanMemRef, copyStartPos.cleanMemRef, size.cleanMemRef)
let gasCost = c.gasCosts[ReturnDataCopy].m_handler(c.memory.len, memPos, len)
c.gasMeter.consumeGas(
2019-04-24 12:26:52 +00:00
gasCost,
2019-03-16 14:42:06 +00:00
reason="returnDataCopy fee")
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
if copyPos + len > c.returnData.len:
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
# TODO Geth additionally checks copyPos + len < 64
# Parity uses a saturating addition
# Yellow paper mentions μs[1] + i are not subject to the 2^256 modulo.
raise newException(OutOfBoundsRead,
"Return data length is not sufficient to satisfy request. Asked \n" &
&"for data from index {copyStartPos} to {copyStartPos + size}. Return data is {c.returnData.len} in \n" &
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
"length")
c.memory.writePaddedResult(c.returnData, memPos, copyPos, len)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
# ##########################################
# 40s: Block Information
op blockhash, inline = true, blockNumber:
## 0x40, Get the hash of one of the 256 most recent complete blocks.
2020-01-16 10:23:51 +00:00
push: c.getBlockHash(blockNumber)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op coinbase, inline = true:
## 0x41, Get the block's beneficiary address.
2020-01-16 10:23:51 +00:00
push: c.getCoinbase()
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op timestamp, inline = true:
## 0x42, Get the block's timestamp.
2020-01-16 10:23:51 +00:00
push: c.getTimestamp()
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op blocknumber, inline = true:
## 0x43, Get the block's number.
2020-01-16 10:23:51 +00:00
push: c.getBlockNumber()
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op difficulty, inline = true:
## 0x44, Get the block's difficulty
2020-01-16 10:23:51 +00:00
push: c.getDifficulty()
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op gasLimit, inline = true:
## 0x45, Get the block's gas limit
2020-01-16 10:23:51 +00:00
push: c.getGasLimit()
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op baseFee, inline = true:
## 0x45, Get the block's gas limit
push: c.getBaseFee()
2019-11-12 12:41:19 +00:00
op chainId, inline = true:
2019-11-11 04:58:58 +00:00
## 0x46, Get current chains EIP-155 unique identifier.
2020-01-16 10:23:51 +00:00
push: c.getChainId()
2019-11-11 04:58:58 +00:00
2019-11-11 05:12:58 +00:00
op selfBalance, inline = true:
## 0x47, Get current contract's balance.
2020-01-16 15:07:04 +00:00
push: c.getBalance(c.msg.contractAddress)
2019-11-11 05:12:58 +00:00
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
# ##########################################
# 50s: Stack, Memory, Storage and Flow Operations
op pop, inline = true:
## 0x50, Remove item from stack.
discard c.stack.popInt()
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op mload, inline = true, memStartPos:
## 0x51, Load word from memory
let memPos = memStartPos.cleanMemRef
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
c.gasMeter.consumeGas(
2022-04-08 04:54:11 +00:00
c.gasCosts[Mload].m_handler(c.memory.len, memPos, 32),
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
reason="MLOAD: GasVeryLow + memory expansion"
)
c.memory.extend(memPos, 32)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
push: c.memory.read(memPos, 32) # TODO, should we convert to native endianness?
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op mstore, inline = true, memStartPos, value:
## 0x52, Save word to memory
let memPos = memStartPos.cleanMemRef
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
c.gasMeter.consumeGas(
2022-04-08 04:54:11 +00:00
c.gasCosts[Mstore].m_handler(c.memory.len, memPos, 32),
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
reason="MSTORE: GasVeryLow + memory expansion"
)
c.memory.extend(memPos, 32)
c.memory.write(memPos, value.toByteArrayBE) # is big-endian correct? Parity/Geth do convert
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op mstore8, inline = true, memStartPos, value:
## 0x53, Save byte to memory
let memPos = memStartPos.cleanMemRef
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
c.gasMeter.consumeGas(
2022-04-08 04:54:11 +00:00
c.gasCosts[Mstore].m_handler(c.memory.len, memPos, 1),
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
reason="MSTORE8: GasVeryLow + memory expansion"
)
c.memory.extend(memPos, 1)
c.memory.write(memPos, [value.toByteArrayBE[31]])
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op sload, inline = true, slot:
## 0x54, Load word from storage.
2020-01-16 14:56:59 +00:00
push: c.getStorage(slot)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2020-07-21 06:15:06 +00:00
when not evmc_enabled:
2022-04-08 04:54:11 +00:00
template sstoreImpl(c: Computation, slot, newValue: UInt256) =
2020-07-21 06:15:06 +00:00
let currentValue {.inject.} = c.getStorage(slot)
2020-01-17 14:49:22 +00:00
2020-07-21 06:15:06 +00:00
let
2020-07-21 13:13:58 +00:00
gasParam = GasParams(kind: Op.Sstore, s_currentValue: currentValue)
2020-07-21 06:15:06 +00:00
(gasCost, gasRefund) = c.gasCosts[Sstore].c_handler(newValue, gasParam)
2020-01-17 14:49:22 +00:00
2020-07-21 06:15:06 +00:00
c.gasMeter.consumeGas(gasCost, &"SSTORE: {c.msg.contractAddress}[{slot}] -> {newValue} ({currentValue})")
2020-01-17 14:49:22 +00:00
2020-07-21 06:15:06 +00:00
if gasRefund > 0:
c.gasMeter.refundGas(gasRefund)
2020-01-17 14:49:22 +00:00
2020-07-21 06:15:06 +00:00
c.vmState.mutateStateDB:
db.setStorage(c.msg.contractAddress, slot, newValue)
2020-01-17 14:49:22 +00:00
when evmc_enabled:
2022-04-08 04:54:11 +00:00
template sstoreEvmc(c: Computation, slot, newValue: UInt256) =
let
currentValue {.inject.} = c.getStorage(slot)
status = c.host.setStorage(c.msg.contractAddress, slot, newValue)
gasParam = GasParams(kind: Op.Sstore, s_status: status)
gasCost = c.gasCosts[Sstore].c_handler(newValue, gasParam)[0]
2020-01-17 14:49:22 +00:00
c.gasMeter.consumeGas(gasCost, &"SSTORE: {c.msg.contractAddress}[{slot}] -> {newValue} ({currentValue})")
2020-01-17 14:49:22 +00:00
op sstore, inline = false, slot, newValue:
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
## 0x55, Save word to storage.
checkInStaticContext(c)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2020-01-17 14:49:22 +00:00
when evmc_enabled:
sstoreEvmc(c, slot, newValue)
else:
sstoreImpl(c, slot, newValue)
2020-07-21 06:15:06 +00:00
when not evmc_enabled:
2022-04-08 04:54:11 +00:00
template sstoreNetGasMeteringImpl(c: Computation, slot, newValue: UInt256) =
2020-07-21 06:15:06 +00:00
let stateDB = c.vmState.readOnlyStateDB
let currentValue {.inject.} = c.getStorage(slot)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2020-07-21 06:15:06 +00:00
let
gasParam = GasParams(kind: Op.Sstore,
s_currentValue: currentValue,
s_originalValue: stateDB.getCommittedStorage(c.msg.contractAddress, slot)
)
(gasCost, gasRefund) = c.gasCosts[Sstore].c_handler(newValue, gasParam)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2020-07-21 06:15:06 +00:00
c.gasMeter.consumeGas(gasCost, &"SSTORE EIP2200: {c.msg.contractAddress}[{slot}] -> {newValue} ({currentValue})")
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2020-07-21 06:15:06 +00:00
if gasRefund != 0:
c.gasMeter.refundGas(gasRefund)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2020-07-21 06:15:06 +00:00
c.vmState.mutateStateDB:
db.setStorage(c.msg.contractAddress, slot, newValue)
2020-01-17 14:49:22 +00:00
op sstoreEIP2200, inline = false, slot, newValue:
checkInStaticContext(c)
const SentryGasEIP2200 = 2300 # Minimum gas required to be present for an SSTORE call, not consumed
if c.gasMeter.gasRemaining <= SentryGasEIP2200:
raise newException(OutOfGas, "Gas not enough to perform EIP2200 SSTORE")
when evmc_enabled:
sstoreEvmc(c, slot, newValue)
else:
2020-03-04 09:32:30 +00:00
sstoreNetGasMeteringImpl(c, slot, newValue)
op sstoreEIP1283, inline = false, slot, newValue:
checkInStaticContext(c)
when evmc_enabled:
sstoreEvmc(c, slot, newValue)
else:
sstoreNetGasMeteringImpl(c, slot, newValue)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
proc jumpImpl(c: Computation, jumpTarget: UInt256) =
if jumpTarget >= c.code.len.u256:
raise newException(InvalidJumpDestination, "Invalid Jump Destination")
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2019-11-13 14:49:39 +00:00
let jt = jumpTarget.truncate(int)
c.code.pc = jt
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
let nextOpcode = c.code.peek
2022-04-08 04:54:11 +00:00
if nextOpcode != JumpDest:
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
raise newException(InvalidJumpDestination, "Invalid Jump Destination")
# TODO: next check seems redundant
if not c.code.isValidOpcode(jt):
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
raise newException(InvalidInstruction, "Jump resulted in invalid instruction")
op jump, inline = true, jumpTarget:
## 0x56, Alter the program counter
jumpImpl(c, jumpTarget)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op jumpI, inline = true, jumpTarget, testedValue:
## 0x57, Conditionally alter the program counter.
if testedValue != 0:
jumpImpl(c, jumpTarget)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op pc, inline = true:
## 0x58, Get the value of the program counter prior to the increment corresponding to this instruction.
push: max(c.code.pc - 1, 0)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op msize, inline = true:
## 0x59, Get the size of active memory in bytes.
push: c.memory.len
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op gas, inline = true:
## 0x5a, Get the amount of available gas, including the corresponding reduction for the cost of this instruction.
push: c.gasMeter.gasRemaining
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op jumpDest, inline = true:
## 0x5b, Mark a valid destination for jumps. This operation has no effect on machine state during execution.
discard
2020-11-25 09:43:34 +00:00
op beginSub, inline = true:
## 0x5c, Marks the entry point to a subroutine
2020-11-25 10:09:10 +00:00
raise newException(OutOfGas, "Abort: Attempt to execute BeginSub opcode")
2020-11-25 09:43:34 +00:00
op returnSub, inline = true:
## 0x5d, Returns control to the caller of a subroutine.
2020-11-25 10:09:10 +00:00
if c.returnStack.len == 0:
raise newException(OutOfGas, "Abort: invalid returnStack during ReturnSub")
# Other than the check that the return stack is not empty, there is no
# need to validate the pc from 'returns', since we only ever push valid
# values onto it via jumpsub.
c.code.pc = c.returnStack.pop()
2020-11-25 10:09:10 +00:00
op jumpSub, inline = true, jumpTarget:
2020-11-25 09:43:34 +00:00
## 0x5e, Transfers control to a subroutine.
2020-11-25 10:09:10 +00:00
if jumpTarget >= c.code.len.u256:
raise newException(InvalidJumpDestination, "JumpSub destination exceeds code len")
2020-11-25 11:23:02 +00:00
let returnPC = c.code.pc
2020-11-25 10:09:10 +00:00
let jt = jumpTarget.truncate(int)
c.code.pc = jt
let nextOpcode = c.code.peek
if nextOpcode != BeginSub:
raise newException(InvalidJumpDestination, "Invalid JumpSub destination")
if c.returnStack.len == 1023:
raise newException(FullStack, "Out of returnStack")
c.returnStack.add returnPC
2020-11-25 10:09:10 +00:00
inc c.code.pc
2020-11-25 09:43:34 +00:00
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
# ##########################################
# 60s & 70s: Push Operations.
# 80s: Duplication Operations
# 90s: Exchange Operations
# a0s: Logging Operations
genPush()
genDup()
genSwap()
genLog()
# ##########################################
# f0s: System operations.
2020-01-31 01:08:44 +00:00
template genCreate(callName: untyped, opCode: Op): untyped =
op callName, inline = false:
checkInStaticContext(c)
let
endowment = c.stack.popInt()
memPos = c.stack.popInt().safeInt
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2020-01-31 01:08:44 +00:00
when opCode == Create:
const callKind = evmcCreate
let memLen {.inject.} = c.stack.peekInt().safeInt
Transaction: EVMC fix, `CREATE2` salt is a 256-bit blob not a number This changes fixes a bug in `CREATE2` ops when used with EVMC. Because it changes the salt type, it affects non-EVMC code as well. The salt was passed through EVMC with the wrong byte order, although this went unnoticed as the Nimbus host flipped the byte order before using it. This was found when running Nimbus with third-party EVM, ["evmone"](https://github.com/ethereum/evmone). There are different ways to remedy this. If treated as a number, Nimbus EVM would byte-flip the value when calling EVMC, then Nimbus host would flip the received value. Finally, it would be flipped a third time when generating the address in `generateSafeAddress`. The first two flips can be eliminated by negotiation (like other numbers), but there would always be one flip. As a bit pattern, Nimbus EVM would flip the same way it does when dealing with hashes on the stack (e.g. with `getBlockHash`). Nimbus host wouldn't flip at all - and when using third-party EVMs there would be no flips in Nimbus. Because this value is not for arithmetic, any bit pattern is valid, and there shouldn't be any flips when using a third-party EVM, the bit-pattern interpretation is favoured. The only flip is done in Nimbus EVM (and might be eliminated in an optimised version). As suggested, we'll define a new "opaque 256 bits" type to hold this value. (Similar to `Hash256`, but the salt isn't necessarily a hash.) Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-06-09 22:32:42 +00:00
const salt: ContractSalt = ZERO_CONTRACTSALT
2020-01-31 01:08:44 +00:00
else:
const callKind = evmcCreate2
let memLen {.inject.} = c.stack.popInt().safeInt
Transaction: EVMC fix, `CREATE2` salt is a 256-bit blob not a number This changes fixes a bug in `CREATE2` ops when used with EVMC. Because it changes the salt type, it affects non-EVMC code as well. The salt was passed through EVMC with the wrong byte order, although this went unnoticed as the Nimbus host flipped the byte order before using it. This was found when running Nimbus with third-party EVM, ["evmone"](https://github.com/ethereum/evmone). There are different ways to remedy this. If treated as a number, Nimbus EVM would byte-flip the value when calling EVMC, then Nimbus host would flip the received value. Finally, it would be flipped a third time when generating the address in `generateSafeAddress`. The first two flips can be eliminated by negotiation (like other numbers), but there would always be one flip. As a bit pattern, Nimbus EVM would flip the same way it does when dealing with hashes on the stack (e.g. with `getBlockHash`). Nimbus host wouldn't flip at all - and when using third-party EVMs there would be no flips in Nimbus. Because this value is not for arithmetic, any bit pattern is valid, and there shouldn't be any flips when using a third-party EVM, the bit-pattern interpretation is favoured. The only flip is done in Nimbus EVM (and might be eliminated in an optimised version). As suggested, we'll define a new "opaque 256 bits" type to hold this value. (Similar to `Hash256`, but the salt isn't necessarily a hash.) Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-06-09 22:32:42 +00:00
let salt = ContractSalt(bytes: c.stack.peekInt().toBytesBE)
2020-01-31 01:08:44 +00:00
c.stack.top(0)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2020-01-31 01:08:44 +00:00
let gasParams = GasParams(kind: Create,
cr_currentMemSize: c.memory.len,
cr_memOffset: memPos,
cr_memLength: memLen
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
)
2020-01-31 01:08:44 +00:00
var gasCost = c.gasCosts[Create].c_handler(1.u256, gasParams).gasCost
when opCode == Create2:
gasCost = gasCost + c.gasCosts[Create2].m_handler(0, 0, memLen)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2020-01-31 01:08:44 +00:00
let reason = &"CREATE: GasCreate + {memLen} * memory expansion"
c.gasMeter.consumeGas(gasCost, reason = reason)
c.memory.extend(memPos, memLen)
c.returnData.setLen(0)
2020-01-15 00:22:09 +00:00
2020-01-31 01:08:44 +00:00
if c.msg.depth >= MaxCallDepth:
debug "Computation Failure", reason = "Stack too deep", maxDepth = MaxCallDepth, depth = c.msg.depth
2019-05-07 03:23:27 +00:00
return
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2020-01-31 01:08:44 +00:00
if endowment != 0:
let senderBalance = c.getBalance(c.msg.contractAddress)
if senderBalance < endowment:
debug "Computation Failure", reason = "Insufficient funds available to transfer", required = endowment, balance = senderBalance
return
var createMsgGas = c.gasMeter.gasRemaining
if c.fork >= FkTangerine:
createMsgGas -= createMsgGas div 64
c.gasMeter.consumeGas(createMsgGas, reason="CREATE")
when evmc_enabled:
EVMC: Small stacks when using EVMC, closes #575 (segfaults) This patch reduces stack space used with EVM in ENABLE_EVMC=1 mode, from 13 MB worst case to 550 kB, a 24x reduction. This completes fixing the "stack problem" and closes #575 (`EVM: Different segmentation faults when running the test suite with EVMC`). It also closes #256 (`recursive EVM call trigger unrecoverable stack overflow`). After this patch, it is possible to re-enable the CI targets which had to be disabled due to #575. This change is also a required precursor for switching over to "nearly EVMC" as the clean and focused Nimbus-internal API between EVM and sync/database processes, and is also key to the use of Chronos `async` in those processes when calling the EVM. (The motivation is the internal interface has to be substantially changed _anyway_ for the parallel sync and database processes, and EVMC turns out to be well-designed and well-suited for this. It provides good separation between modules, and suits our needs better than our other current interface. Might as well use a good one designed by someone else. EVMC is 98% done in Nimbus thanks to great work done before by @jangko, and we can use Nimbus-specific extensions where we need flexibility, including for performance. Being aligned with the ecosystem is a useful bonus feature.) All tests below were run on Ubuntu 20.04 LTS server, x86-64. This matches one of the targets that has been disabled for a while in CI in EVMC mode due to stack overflow crashing the tests, so it's a good choice. Measurements before =================== Testing commit `e76e0144 2021-04-22 11:29:42 +0700 add submodules: graphql and toml-serialization`. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 38416 depthHigh 3 ... Stack range 13074720 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 13.07 MB of stack to run, and so crash with the default stack limit on Ubuntu Server 20.04 (8MB). Exactly 12768 bytes per EVM call stack frame. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 14384 depthHigh 2 ... Stack range 3495456 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 3709600 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 7831600 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 7.83MB of stack to run. About 7648 bytes per EVM call stack frame. It _only just_ avoids crashing with the default Ubuntu Server stack limit of 8 MB. However, it still crashes on Windows x86-64, which is why the Windows CI EVMC target is currently disabled. On Linux where this passes, this is so borderline that it affects work and testing of the complex storage code, because that's called from the EVM. Also, this greatly exceeds the default thread stack size. Measurements after ================== $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 1936 depthHigh 3 ... Stack range 556272 depthHigh 1022 Stack range 556512 depthHigh 1023 Stack range 556816 depthHigh 1023 Stack range 557056 depthHigh 1024 Stack range 557360 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 1392 depthHigh 2 ... Stack range 248912 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 264144 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 557360 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stStaticCall/static_CallRecursiveBombPreCall.json For both tests, a satisfying *544 bytes* per EVM call stack frame, and EVM takes less than 600 kB total. With other overheads, both tests run in 600 kB stack total at maximum EVM depth. We must add some headroom on this for database activity called from the EVM, and different compile targets. But it means the EVM itself is no longer a stack burden. This is much smaller than the default thread stack size on Linux (2MB), with plenty of margin. (Just fyi, it isn't smaller than a _small_ thread stack on Linux from a long time ago (128kB), and some small embedded C targets.) This size is well suited to running EVMs in threads. Further reduction ================= This patch solves the stack problem. Windows and Linux 64-bit EVMC CI targets can be re-enabled, and there is no longer a problem with stack usage. We can reduce further to ~340 bytes per frame and 350 kB total, while still complying with EVMC. But as this involves changing how errors are handled to comply fully with EVMC, and removing `dispose` calls, it's not worth doing now while there are other EVMC changes in progress that will have the same effect. A Nimbus-specific extension will allow us to avoid recursion with EVMC anyway, bringing bytes per frame to zero. We need the extension anyway, to support Chronos `async` which parallel transaction processing is built around. Interop with non-Nimbus over EVMC won't let us avoid recursion, but then we can't control the stack frame size either. To prevent stack overflow in interop I anticipate using (this method in Aleth) [https://github.com/ethereum/aleth/blob/6e96ce34e3f131e2d42f3cb00741b54e05ab029d/libethereum/ExtVM.cpp#L61]. Smoke test other versions of GCC and Clang/LLVM =============================================== As all builds including Windows use GCC or Apple's Clang/LLVM, this is just to verify we're in the right ballpark on all targets. I've only checked `x86_64` though, not 32-bit, and not ARM. It's interesting to see GCC 10 uses less stack. This is because it optimises `struct` returns better, sometimes skipping an intermediate copy. Here it benefits the EVMC API, but I found GCC 10 also improves the larger stack usage of the rest of `nimbus-eth1` as well. Apple clang 12.0.0 (clang-1200.0.26.2) on MacOS 10.15: - 544 bytes per EVM call stack frame GCC 10.3.0 (Ubuntu 10.3.0-1ubuntu1) on Ubuntu 21.04: - 464 bytes per EVM call stack frame GCC 10.2.0 (Ubuntu 10.2.0-5ubuntu1~20.04) on Ubuntu 20.04 LTS: - 464 bytes per EVM call stack frame GCC 11.0.1 20210417 (experimental; Ubuntu 11-20210417-1ubuntu1) on Ubuntu 21.04: - 8 bytes per EVM call stack frame GCC 9.3.0 (Ubuntu 9.3.0-17ubuntu1~20.04) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 8.4.0 (Ubuntu 8.4.0-3ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 7.5.0 (Ubuntu 7.5.0-6ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 9.2.1 20191008 (Ubuntu 9.2.1-9ubuntu2) on Ubuntu 19.10: - 528 bytes per EVM call stack frame Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-04-26 15:01:00 +00:00
let msg = new(nimbus_message)
msg[] = nimbus_message(
2020-01-31 01:08:44 +00:00
kind: callKind.evmc_call_kind,
depth: (c.msg.depth + 1).int32,
gas: createMsgGas,
sender: c.msg.contractAddress,
input_data: c.memory.readPtr(memPos),
input_size: memLen.uint,
value: toEvmc(endowment),
Transaction: EVMC fix, `CREATE2` salt is a 256-bit blob not a number This changes fixes a bug in `CREATE2` ops when used with EVMC. Because it changes the salt type, it affects non-EVMC code as well. The salt was passed through EVMC with the wrong byte order, although this went unnoticed as the Nimbus host flipped the byte order before using it. This was found when running Nimbus with third-party EVM, ["evmone"](https://github.com/ethereum/evmone). There are different ways to remedy this. If treated as a number, Nimbus EVM would byte-flip the value when calling EVMC, then Nimbus host would flip the received value. Finally, it would be flipped a third time when generating the address in `generateSafeAddress`. The first two flips can be eliminated by negotiation (like other numbers), but there would always be one flip. As a bit pattern, Nimbus EVM would flip the same way it does when dealing with hashes on the stack (e.g. with `getBlockHash`). Nimbus host wouldn't flip at all - and when using third-party EVMs there would be no flips in Nimbus. Because this value is not for arithmetic, any bit pattern is valid, and there shouldn't be any flips when using a third-party EVM, the bit-pattern interpretation is favoured. The only flip is done in Nimbus EVM (and might be eliminated in an optimised version). As suggested, we'll define a new "opaque 256 bits" type to hold this value. (Similar to `Hash256`, but the salt isn't necessarily a hash.) Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-06-09 22:32:42 +00:00
create2_salt: toEvmc(salt),
2020-01-31 01:08:44 +00:00
)
EVMC: Small stacks when using EVMC, closes #575 (segfaults) This patch reduces stack space used with EVM in ENABLE_EVMC=1 mode, from 13 MB worst case to 550 kB, a 24x reduction. This completes fixing the "stack problem" and closes #575 (`EVM: Different segmentation faults when running the test suite with EVMC`). It also closes #256 (`recursive EVM call trigger unrecoverable stack overflow`). After this patch, it is possible to re-enable the CI targets which had to be disabled due to #575. This change is also a required precursor for switching over to "nearly EVMC" as the clean and focused Nimbus-internal API between EVM and sync/database processes, and is also key to the use of Chronos `async` in those processes when calling the EVM. (The motivation is the internal interface has to be substantially changed _anyway_ for the parallel sync and database processes, and EVMC turns out to be well-designed and well-suited for this. It provides good separation between modules, and suits our needs better than our other current interface. Might as well use a good one designed by someone else. EVMC is 98% done in Nimbus thanks to great work done before by @jangko, and we can use Nimbus-specific extensions where we need flexibility, including for performance. Being aligned with the ecosystem is a useful bonus feature.) All tests below were run on Ubuntu 20.04 LTS server, x86-64. This matches one of the targets that has been disabled for a while in CI in EVMC mode due to stack overflow crashing the tests, so it's a good choice. Measurements before =================== Testing commit `e76e0144 2021-04-22 11:29:42 +0700 add submodules: graphql and toml-serialization`. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 38416 depthHigh 3 ... Stack range 13074720 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 13.07 MB of stack to run, and so crash with the default stack limit on Ubuntu Server 20.04 (8MB). Exactly 12768 bytes per EVM call stack frame. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 14384 depthHigh 2 ... Stack range 3495456 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 3709600 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 7831600 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 7.83MB of stack to run. About 7648 bytes per EVM call stack frame. It _only just_ avoids crashing with the default Ubuntu Server stack limit of 8 MB. However, it still crashes on Windows x86-64, which is why the Windows CI EVMC target is currently disabled. On Linux where this passes, this is so borderline that it affects work and testing of the complex storage code, because that's called from the EVM. Also, this greatly exceeds the default thread stack size. Measurements after ================== $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 1936 depthHigh 3 ... Stack range 556272 depthHigh 1022 Stack range 556512 depthHigh 1023 Stack range 556816 depthHigh 1023 Stack range 557056 depthHigh 1024 Stack range 557360 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 1392 depthHigh 2 ... Stack range 248912 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 264144 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 557360 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stStaticCall/static_CallRecursiveBombPreCall.json For both tests, a satisfying *544 bytes* per EVM call stack frame, and EVM takes less than 600 kB total. With other overheads, both tests run in 600 kB stack total at maximum EVM depth. We must add some headroom on this for database activity called from the EVM, and different compile targets. But it means the EVM itself is no longer a stack burden. This is much smaller than the default thread stack size on Linux (2MB), with plenty of margin. (Just fyi, it isn't smaller than a _small_ thread stack on Linux from a long time ago (128kB), and some small embedded C targets.) This size is well suited to running EVMs in threads. Further reduction ================= This patch solves the stack problem. Windows and Linux 64-bit EVMC CI targets can be re-enabled, and there is no longer a problem with stack usage. We can reduce further to ~340 bytes per frame and 350 kB total, while still complying with EVMC. But as this involves changing how errors are handled to comply fully with EVMC, and removing `dispose` calls, it's not worth doing now while there are other EVMC changes in progress that will have the same effect. A Nimbus-specific extension will allow us to avoid recursion with EVMC anyway, bringing bytes per frame to zero. We need the extension anyway, to support Chronos `async` which parallel transaction processing is built around. Interop with non-Nimbus over EVMC won't let us avoid recursion, but then we can't control the stack frame size either. To prevent stack overflow in interop I anticipate using (this method in Aleth) [https://github.com/ethereum/aleth/blob/6e96ce34e3f131e2d42f3cb00741b54e05ab029d/libethereum/ExtVM.cpp#L61]. Smoke test other versions of GCC and Clang/LLVM =============================================== As all builds including Windows use GCC or Apple's Clang/LLVM, this is just to verify we're in the right ballpark on all targets. I've only checked `x86_64` though, not 32-bit, and not ARM. It's interesting to see GCC 10 uses less stack. This is because it optimises `struct` returns better, sometimes skipping an intermediate copy. Here it benefits the EVMC API, but I found GCC 10 also improves the larger stack usage of the rest of `nimbus-eth1` as well. Apple clang 12.0.0 (clang-1200.0.26.2) on MacOS 10.15: - 544 bytes per EVM call stack frame GCC 10.3.0 (Ubuntu 10.3.0-1ubuntu1) on Ubuntu 21.04: - 464 bytes per EVM call stack frame GCC 10.2.0 (Ubuntu 10.2.0-5ubuntu1~20.04) on Ubuntu 20.04 LTS: - 464 bytes per EVM call stack frame GCC 11.0.1 20210417 (experimental; Ubuntu 11-20210417-1ubuntu1) on Ubuntu 21.04: - 8 bytes per EVM call stack frame GCC 9.3.0 (Ubuntu 9.3.0-17ubuntu1~20.04) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 8.4.0 (Ubuntu 8.4.0-3ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 7.5.0 (Ubuntu 7.5.0-6ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 9.2.1 20191008 (Ubuntu 9.2.1-9ubuntu2) on Ubuntu 19.10: - 528 bytes per EVM call stack frame Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-04-26 15:01:00 +00:00
c.chainTo(msg):
c.gasMeter.returnGas(c.res.gas_left)
if c.res.status_code == EVMC_SUCCESS:
c.stack.top(c.res.create_address)
EVM: `writeContract` fixes, never return contract code as `RETURNDATA` This fixes #867 "EIP-170 related consensus error at Goerli block 5080941", and equivalent on other networks. This combines a change on the EVM-caller side with an EVM-side change from @jangko 6548ff98 "fixes CREATE/CREATE2's `returndata` bug", making the caller EVM ignore any data except from `REVERT`. Either change works by itself. The reason for both is to ensure we definitely comply with ambiguous EVMC expectations from either side of that boundary, and it makes the internal API clearer. As well as fixing a specific consensus issue, there are some other EVM logic changes too: Refactored `writeContract`, how `RETURNDATA` is handled inside the EVM, and changed behaviour with quirks before EIP-2 (Homestead). The fix allows sync to pass block 5080941 on Goerli, and probably equivalent on other networks. Here's a trace at batch 5080897..5081088: ``` TRC 2021-10-01 21:18:12.883+01:00 Persisting blocks file=persist_blocks.nim:43 fromBlock=5080897 toBlock=5081088 ... DBG 2021-10-01 21:18:13.270+01:00 Contract code size exceeds EIP170 topics="vm computation" file=computation.nim:236 limit=24577 actual=31411 DBG 2021-10-01 21:18:13.271+01:00 gasUsed neq cumulativeGasUsed file=process_block.nim:68 block=5080941/0A3537BC5BDFC637349E1C77D9648F2F65E2BF973ABF7956618F854B769DF626 gasUsed=3129669 cumulativeGasUsed=3132615 TRC 2021-10-01 21:18:13.271+01:00 peer disconnected file=blockchain_sync.nim:407 peer=<IP:PORT> ``` Although it says "Contract code size" and "gasUsed", this bug is more general than either contract size or gas. It's due to incorrect behaviour of EVM instructions `RETURNDATA` and `RETURNDATASIZE`. Sometimes when `writeContract` decides to reject writing the contract for any of several reasons (for example just insufficient gas), the unwritten contract code was being used as the "return data", and given to the caller. If the caller used `RETURNDATA` or `RETURNDATASIZE` ops, those incorrectly reported the contract code that didn't get written. EIP-211 (https://eips.ethereum.org/EIPS/eip-211) describes `RETURNDATA`: > "`CREATE` and `CREATE2` are considered to return the empty buffer in the > success case and the failure data in the failure case". The language is ambiguous. In fact "failure case" means when the contract uses `REVERT` to finish. It doesn't mean other failures like out of gas, EIP-170 limit, EIP-3541, etc. To be thorough, and to ensure we always do the right thing with real EVMC when that's finalised, this patch fixes the `RETURNDATA` issue in two places, either of which make Goerli block 5080941 pass. `writeContract` has been refactored to be caller, and so has where it's called. It sets an error in the usual way if contract writing is rejected -- that's anticipating EVMC, where we'll use different error codes later. Overall four behaviour changes: 1. On the callee side, it doesn't set `c.outputData` except for `REVERT`. 2. On the caller side, it doesn't read `child.outputData` except for `REVERT`. 3. There was a bug in processing before Homestead fork (EIP-2). We did not match the spec or other implementations; now we do. When there's insufficient gas, before Homestead it's treated as success but with an empty contract. https://github.com/ethereum/pyethereum/blob/d117c8f3fd93359fc641fd850fa799436f7c43b5/ethereum/processblock.py#L304 https://github.com/ethereum/go-ethereum/blob/401354976bb4/core/vm/instructions.go#L586 4. The Byzantium check has been removed, as it's unnecessary. Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-12-02 19:44:51 +00:00
elif c.res.status_code == EVMC_REVERT:
# From create, only use `outputData` if child returned with `REVERT`.
c.returnData = @(makeOpenArray(c.res.outputData, c.res.outputSize.int))
EVMC: Small stacks when using EVMC, closes #575 (segfaults) This patch reduces stack space used with EVM in ENABLE_EVMC=1 mode, from 13 MB worst case to 550 kB, a 24x reduction. This completes fixing the "stack problem" and closes #575 (`EVM: Different segmentation faults when running the test suite with EVMC`). It also closes #256 (`recursive EVM call trigger unrecoverable stack overflow`). After this patch, it is possible to re-enable the CI targets which had to be disabled due to #575. This change is also a required precursor for switching over to "nearly EVMC" as the clean and focused Nimbus-internal API between EVM and sync/database processes, and is also key to the use of Chronos `async` in those processes when calling the EVM. (The motivation is the internal interface has to be substantially changed _anyway_ for the parallel sync and database processes, and EVMC turns out to be well-designed and well-suited for this. It provides good separation between modules, and suits our needs better than our other current interface. Might as well use a good one designed by someone else. EVMC is 98% done in Nimbus thanks to great work done before by @jangko, and we can use Nimbus-specific extensions where we need flexibility, including for performance. Being aligned with the ecosystem is a useful bonus feature.) All tests below were run on Ubuntu 20.04 LTS server, x86-64. This matches one of the targets that has been disabled for a while in CI in EVMC mode due to stack overflow crashing the tests, so it's a good choice. Measurements before =================== Testing commit `e76e0144 2021-04-22 11:29:42 +0700 add submodules: graphql and toml-serialization`. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 38416 depthHigh 3 ... Stack range 13074720 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 13.07 MB of stack to run, and so crash with the default stack limit on Ubuntu Server 20.04 (8MB). Exactly 12768 bytes per EVM call stack frame. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 14384 depthHigh 2 ... Stack range 3495456 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 3709600 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 7831600 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 7.83MB of stack to run. About 7648 bytes per EVM call stack frame. It _only just_ avoids crashing with the default Ubuntu Server stack limit of 8 MB. However, it still crashes on Windows x86-64, which is why the Windows CI EVMC target is currently disabled. On Linux where this passes, this is so borderline that it affects work and testing of the complex storage code, because that's called from the EVM. Also, this greatly exceeds the default thread stack size. Measurements after ================== $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 1936 depthHigh 3 ... Stack range 556272 depthHigh 1022 Stack range 556512 depthHigh 1023 Stack range 556816 depthHigh 1023 Stack range 557056 depthHigh 1024 Stack range 557360 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 1392 depthHigh 2 ... Stack range 248912 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 264144 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 557360 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stStaticCall/static_CallRecursiveBombPreCall.json For both tests, a satisfying *544 bytes* per EVM call stack frame, and EVM takes less than 600 kB total. With other overheads, both tests run in 600 kB stack total at maximum EVM depth. We must add some headroom on this for database activity called from the EVM, and different compile targets. But it means the EVM itself is no longer a stack burden. This is much smaller than the default thread stack size on Linux (2MB), with plenty of margin. (Just fyi, it isn't smaller than a _small_ thread stack on Linux from a long time ago (128kB), and some small embedded C targets.) This size is well suited to running EVMs in threads. Further reduction ================= This patch solves the stack problem. Windows and Linux 64-bit EVMC CI targets can be re-enabled, and there is no longer a problem with stack usage. We can reduce further to ~340 bytes per frame and 350 kB total, while still complying with EVMC. But as this involves changing how errors are handled to comply fully with EVMC, and removing `dispose` calls, it's not worth doing now while there are other EVMC changes in progress that will have the same effect. A Nimbus-specific extension will allow us to avoid recursion with EVMC anyway, bringing bytes per frame to zero. We need the extension anyway, to support Chronos `async` which parallel transaction processing is built around. Interop with non-Nimbus over EVMC won't let us avoid recursion, but then we can't control the stack frame size either. To prevent stack overflow in interop I anticipate using (this method in Aleth) [https://github.com/ethereum/aleth/blob/6e96ce34e3f131e2d42f3cb00741b54e05ab029d/libethereum/ExtVM.cpp#L61]. Smoke test other versions of GCC and Clang/LLVM =============================================== As all builds including Windows use GCC or Apple's Clang/LLVM, this is just to verify we're in the right ballpark on all targets. I've only checked `x86_64` though, not 32-bit, and not ARM. It's interesting to see GCC 10 uses less stack. This is because it optimises `struct` returns better, sometimes skipping an intermediate copy. Here it benefits the EVMC API, but I found GCC 10 also improves the larger stack usage of the rest of `nimbus-eth1` as well. Apple clang 12.0.0 (clang-1200.0.26.2) on MacOS 10.15: - 544 bytes per EVM call stack frame GCC 10.3.0 (Ubuntu 10.3.0-1ubuntu1) on Ubuntu 21.04: - 464 bytes per EVM call stack frame GCC 10.2.0 (Ubuntu 10.2.0-5ubuntu1~20.04) on Ubuntu 20.04 LTS: - 464 bytes per EVM call stack frame GCC 11.0.1 20210417 (experimental; Ubuntu 11-20210417-1ubuntu1) on Ubuntu 21.04: - 8 bytes per EVM call stack frame GCC 9.3.0 (Ubuntu 9.3.0-17ubuntu1~20.04) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 8.4.0 (Ubuntu 8.4.0-3ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 7.5.0 (Ubuntu 7.5.0-6ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 9.2.1 20191008 (Ubuntu 9.2.1-9ubuntu2) on Ubuntu 19.10: - 528 bytes per EVM call stack frame Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-04-26 15:01:00 +00:00
if not c.res.release.isNil:
c.res.release(c.res)
2020-01-31 01:08:44 +00:00
else:
let childMsg = Message(
kind: callKind,
depth: c.msg.depth + 1,
gas: createMsgGas,
sender: c.msg.contractAddress,
value: endowment,
data: c.memory.read(memPos, memLen)
)
var child = newComputation(c.vmState, childMsg, salt)
EVM: Small patch that reduces EVM stack usage to almost nothing There's been a lot of talk about the Nimbus EVM "stack problem". I think we assumed changing it would require big changes to the interpreter code, touching a lot of functions. It turned out to be a low hanging fruit. This patch solves the stack problem, but hardly touches anything. The change in EVM stack memory is from 13 MB worst case to just 48 kB, a 250x reduction. I've been doing work on the database/storage/trie code. While looking at the API between the EVM and the database/storage/trie, this stack patch stood out and made itself obvious. As it's tiny, rather than more talk, here it is. Note: This patch is intentionally small, non-invasive, and hopefully easy to understand, so that it doesn't conflict with other work done on the EVM, and can easily be grafted into any other EVM structure. Motivation ========== - We run out of space and crash on some targets, unless the stack limit is raised above its default. Surprise segmentation faults are unhelpful. - Some CI targets have been disabled for months due to this. - Because usage borders on the system limits, when working on database/storage/trie/sync code (called from the EVM), segmentation faults occur and are misleading. They cause lost time due to thinking there's a crash bug in the code being worked on, when there's nothing wrong with it. - Sometimes unrelated, trivial code changes elsewhere trigger CI test failures. It looks like abrupt termination. A simple, recent patch was crashing in `make test` even though it was a trivial refactor. Turns out it pushed the stack over the edge. - A large stack has to be scanned by the Nim garbage collector sometimes. Larger stack means slower GC and memory allocation. - The structure of this small patch suggests how to weave async into the EVM with almost no changes to the EVM, and no async transformation overhead. - The patch seemed obvious when working on the API between EVM and storage. Measurements before =================== All these tests were run on Ubuntu 20.04 server, x86-64. This is one of the targets that has been disabled for a while in CI in EVMC mode due to crashing, and excessive stack usage is the cause. Testing commit 0c34a8e3 `2021-04-08 17:46:00 +0200 CI: use MSYS2 on Windows`. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 38496 depthHigh 3 ... Stack range 13140272 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 13.14 MB of stack to run, and so crash with the default stack limit on Ubuntu Server 20.04 (8MB). Exactly 12832 bytes per EVM call stack frame. It's interesting to see some stack frames take a bit more. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 15488 depthHigh 2 ... Stack range 3539312 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 3756144 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 7929968 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 7.92MB of stack to run. About 7264 bytes per EVM call stack frame. It _only just_ avoids crashing with the default Ubuntu Server stack limit of 8 MB. However, it still crashes on Windows x86-64, which is why the CI target is currently disabled. On Linux where this passes, this is so borderline that it affects work and testing of storage and sync code, because that's called from the EVM. Which was a motivation for dealing with the stack instead of letting this linger. Also, this stack greatly exceeds the default thread stack size. $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 33216 depthHigh 3 ... Stack range 11338032 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 11.33 MB stack to run, and so crash with a default stack limit of 8MB. Exactly 11072 bytes per EVM call stack frame. It's interesting to see some stack frames take a bit more. $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 10224 depthHigh 2 ... Stack range 2471760 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 2623184 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 5537824 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 5.54 MB of stack to run, and avoid crashing on with a default stack limit of 8 MB. About 5408 bytes per EVM call stack frame. However, this is uncomfortably close to the limit, as the stack frame size is sensitive to changes in the code. Also, this stack greatly exceeds the default thread stack size. Measurements after ================== (This patch doesn't address EVMC mode, which is not our default. EVMC stack usage remains about the same. EVMC mode is addressed in another tiny patch.) $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 80 # Because we can! 80k stack. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 496 depthHigh 3 ... Stack range 49504 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 72 # Because we can! 72k stack. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 448 depthHigh 2 ... Stack range 22288 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 23632 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 49504 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json For both tests, a satisfying *48 bytes* per EVM call stack frame, and EVM takes not much more than 48 kB. With other overheads, both tests run in 80 kB stack total at maximum EVM depth. We must add some headroom on this for database activity called from the EVM, and different compile targets. But it means the EVM itself is no longer a stack burden. This is much smaller than the default thread stack size on Linux (2MB), with plenty of margin. It's even smaller than Linux from a long time ago (128kB), and some small embedded C targets. (Just fyi, though, some JVM environments allocated just 32 kB to thread stacks.) This size is also well suited to running EVMs in threads, if that's useful. Subtle exception handling and `dispose` ======================================= It is important that each `snapshot` has a corresponding `dispose` in the event of an exception being raised. This code does do that, but in a subtle way. The pair of functions `execCallOrCreate` and `execCallOrCreateAux` are equivalent to the following code, where you can see `dispose` more clearly: proc execCallOrCreate*(c: Computation) = defer: c.dispose() if c.beforeExec(): return c.executeOpcodes() while not c.continuation.isNil: c.child.execCallOrCreate() c.child = nil (c.continuation)() c.executeOpcodes() c.afterExec() That works fine, but only reduces the stack used to 300-700 kB instead of 48 kB. To get lower we split the above into separate `execCallOrCreate` and `execCallOrCreateAux`. Only the outermost has `defer`, and instead of handling one level, it walks the entire `c.parent` chain calling `dispose` if needed. The inner one avoids `defer`, which greatly reduces the size of its stackframe. `c` is a `var` parameter, at each level of recursion. So the outermost proc sees the temporary changes made by all inner calls. This is why `c` is updated and the `c.parent` chain is maintained at each step. Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-04-12 17:06:31 +00:00
c.chainTo(child):
if not child.shouldBurnGas:
c.gasMeter.returnGas(child.gasMeter.gasRemaining)
if child.isSuccess:
c.merge(child)
c.stack.top child.msg.contractAddress
EVM: `writeContract` fixes, never return contract code as `RETURNDATA` This fixes #867 "EIP-170 related consensus error at Goerli block 5080941", and equivalent on other networks. This combines a change on the EVM-caller side with an EVM-side change from @jangko 6548ff98 "fixes CREATE/CREATE2's `returndata` bug", making the caller EVM ignore any data except from `REVERT`. Either change works by itself. The reason for both is to ensure we definitely comply with ambiguous EVMC expectations from either side of that boundary, and it makes the internal API clearer. As well as fixing a specific consensus issue, there are some other EVM logic changes too: Refactored `writeContract`, how `RETURNDATA` is handled inside the EVM, and changed behaviour with quirks before EIP-2 (Homestead). The fix allows sync to pass block 5080941 on Goerli, and probably equivalent on other networks. Here's a trace at batch 5080897..5081088: ``` TRC 2021-10-01 21:18:12.883+01:00 Persisting blocks file=persist_blocks.nim:43 fromBlock=5080897 toBlock=5081088 ... DBG 2021-10-01 21:18:13.270+01:00 Contract code size exceeds EIP170 topics="vm computation" file=computation.nim:236 limit=24577 actual=31411 DBG 2021-10-01 21:18:13.271+01:00 gasUsed neq cumulativeGasUsed file=process_block.nim:68 block=5080941/0A3537BC5BDFC637349E1C77D9648F2F65E2BF973ABF7956618F854B769DF626 gasUsed=3129669 cumulativeGasUsed=3132615 TRC 2021-10-01 21:18:13.271+01:00 peer disconnected file=blockchain_sync.nim:407 peer=<IP:PORT> ``` Although it says "Contract code size" and "gasUsed", this bug is more general than either contract size or gas. It's due to incorrect behaviour of EVM instructions `RETURNDATA` and `RETURNDATASIZE`. Sometimes when `writeContract` decides to reject writing the contract for any of several reasons (for example just insufficient gas), the unwritten contract code was being used as the "return data", and given to the caller. If the caller used `RETURNDATA` or `RETURNDATASIZE` ops, those incorrectly reported the contract code that didn't get written. EIP-211 (https://eips.ethereum.org/EIPS/eip-211) describes `RETURNDATA`: > "`CREATE` and `CREATE2` are considered to return the empty buffer in the > success case and the failure data in the failure case". The language is ambiguous. In fact "failure case" means when the contract uses `REVERT` to finish. It doesn't mean other failures like out of gas, EIP-170 limit, EIP-3541, etc. To be thorough, and to ensure we always do the right thing with real EVMC when that's finalised, this patch fixes the `RETURNDATA` issue in two places, either of which make Goerli block 5080941 pass. `writeContract` has been refactored to be caller, and so has where it's called. It sets an error in the usual way if contract writing is rejected -- that's anticipating EVMC, where we'll use different error codes later. Overall four behaviour changes: 1. On the callee side, it doesn't set `c.outputData` except for `REVERT`. 2. On the caller side, it doesn't read `child.outputData` except for `REVERT`. 3. There was a bug in processing before Homestead fork (EIP-2). We did not match the spec or other implementations; now we do. When there's insufficient gas, before Homestead it's treated as success but with an empty contract. https://github.com/ethereum/pyethereum/blob/d117c8f3fd93359fc641fd850fa799436f7c43b5/ethereum/processblock.py#L304 https://github.com/ethereum/go-ethereum/blob/401354976bb4/core/vm/instructions.go#L586 4. The Byzantium check has been removed, as it's unnecessary. Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-12-02 19:44:51 +00:00
elif not child.error.burnsGas: # Means return was `REVERT`.
# From create, only use `outputData` if child returned with `REVERT`.
EVM: Small patch that reduces EVM stack usage to almost nothing There's been a lot of talk about the Nimbus EVM "stack problem". I think we assumed changing it would require big changes to the interpreter code, touching a lot of functions. It turned out to be a low hanging fruit. This patch solves the stack problem, but hardly touches anything. The change in EVM stack memory is from 13 MB worst case to just 48 kB, a 250x reduction. I've been doing work on the database/storage/trie code. While looking at the API between the EVM and the database/storage/trie, this stack patch stood out and made itself obvious. As it's tiny, rather than more talk, here it is. Note: This patch is intentionally small, non-invasive, and hopefully easy to understand, so that it doesn't conflict with other work done on the EVM, and can easily be grafted into any other EVM structure. Motivation ========== - We run out of space and crash on some targets, unless the stack limit is raised above its default. Surprise segmentation faults are unhelpful. - Some CI targets have been disabled for months due to this. - Because usage borders on the system limits, when working on database/storage/trie/sync code (called from the EVM), segmentation faults occur and are misleading. They cause lost time due to thinking there's a crash bug in the code being worked on, when there's nothing wrong with it. - Sometimes unrelated, trivial code changes elsewhere trigger CI test failures. It looks like abrupt termination. A simple, recent patch was crashing in `make test` even though it was a trivial refactor. Turns out it pushed the stack over the edge. - A large stack has to be scanned by the Nim garbage collector sometimes. Larger stack means slower GC and memory allocation. - The structure of this small patch suggests how to weave async into the EVM with almost no changes to the EVM, and no async transformation overhead. - The patch seemed obvious when working on the API between EVM and storage. Measurements before =================== All these tests were run on Ubuntu 20.04 server, x86-64. This is one of the targets that has been disabled for a while in CI in EVMC mode due to crashing, and excessive stack usage is the cause. Testing commit 0c34a8e3 `2021-04-08 17:46:00 +0200 CI: use MSYS2 on Windows`. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 38496 depthHigh 3 ... Stack range 13140272 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 13.14 MB of stack to run, and so crash with the default stack limit on Ubuntu Server 20.04 (8MB). Exactly 12832 bytes per EVM call stack frame. It's interesting to see some stack frames take a bit more. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 15488 depthHigh 2 ... Stack range 3539312 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 3756144 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 7929968 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 7.92MB of stack to run. About 7264 bytes per EVM call stack frame. It _only just_ avoids crashing with the default Ubuntu Server stack limit of 8 MB. However, it still crashes on Windows x86-64, which is why the CI target is currently disabled. On Linux where this passes, this is so borderline that it affects work and testing of storage and sync code, because that's called from the EVM. Which was a motivation for dealing with the stack instead of letting this linger. Also, this stack greatly exceeds the default thread stack size. $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 33216 depthHigh 3 ... Stack range 11338032 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 11.33 MB stack to run, and so crash with a default stack limit of 8MB. Exactly 11072 bytes per EVM call stack frame. It's interesting to see some stack frames take a bit more. $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 10224 depthHigh 2 ... Stack range 2471760 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 2623184 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 5537824 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 5.54 MB of stack to run, and avoid crashing on with a default stack limit of 8 MB. About 5408 bytes per EVM call stack frame. However, this is uncomfortably close to the limit, as the stack frame size is sensitive to changes in the code. Also, this stack greatly exceeds the default thread stack size. Measurements after ================== (This patch doesn't address EVMC mode, which is not our default. EVMC stack usage remains about the same. EVMC mode is addressed in another tiny patch.) $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 80 # Because we can! 80k stack. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 496 depthHigh 3 ... Stack range 49504 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 72 # Because we can! 72k stack. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 448 depthHigh 2 ... Stack range 22288 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 23632 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 49504 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json For both tests, a satisfying *48 bytes* per EVM call stack frame, and EVM takes not much more than 48 kB. With other overheads, both tests run in 80 kB stack total at maximum EVM depth. We must add some headroom on this for database activity called from the EVM, and different compile targets. But it means the EVM itself is no longer a stack burden. This is much smaller than the default thread stack size on Linux (2MB), with plenty of margin. It's even smaller than Linux from a long time ago (128kB), and some small embedded C targets. (Just fyi, though, some JVM environments allocated just 32 kB to thread stacks.) This size is also well suited to running EVMs in threads, if that's useful. Subtle exception handling and `dispose` ======================================= It is important that each `snapshot` has a corresponding `dispose` in the event of an exception being raised. This code does do that, but in a subtle way. The pair of functions `execCallOrCreate` and `execCallOrCreateAux` are equivalent to the following code, where you can see `dispose` more clearly: proc execCallOrCreate*(c: Computation) = defer: c.dispose() if c.beforeExec(): return c.executeOpcodes() while not c.continuation.isNil: c.child.execCallOrCreate() c.child = nil (c.continuation)() c.executeOpcodes() c.afterExec() That works fine, but only reduces the stack used to 300-700 kB instead of 48 kB. To get lower we split the above into separate `execCallOrCreate` and `execCallOrCreateAux`. Only the outermost has `defer`, and instead of handling one level, it walks the entire `c.parent` chain calling `dispose` if needed. The inner one avoids `defer`, which greatly reduces the size of its stackframe. `c` is a `var` parameter, at each level of recursion. So the outermost proc sees the temporary changes made by all inner calls. This is why `c` is updated and the `c.parent` chain is maintained at each step. Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-04-12 17:06:31 +00:00
c.returnData = child.output
2019-05-07 03:23:27 +00:00
genCreate(create, Create)
genCreate(create2, Create2)
2019-02-20 12:31:17 +00:00
2020-02-03 05:37:33 +00:00
proc callParams(c: Computation): (UInt256, UInt256, EthAddress, EthAddress, CallKind, int, int, int, int, MsgFlags) =
let gas = c.stack.popInt()
2020-02-03 05:37:33 +00:00
let destination = c.stack.popAddress()
2020-01-31 11:23:33 +00:00
let value = c.stack.popInt()
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
result = (gas,
value,
2020-02-03 05:37:33 +00:00
destination,
c.msg.contractAddress, # sender
evmcCall,
2020-01-31 11:23:33 +00:00
c.stack.popInt().cleanMemRef,
c.stack.popInt().cleanMemRef,
c.stack.popInt().cleanMemRef,
c.stack.popInt().cleanMemRef,
c.msg.flags)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2020-02-03 05:37:33 +00:00
proc callCodeParams(c: Computation): (UInt256, UInt256, EthAddress, EthAddress, CallKind, int, int, int, int, MsgFlags) =
let gas = c.stack.popInt()
2020-02-03 05:37:33 +00:00
let destination = c.stack.popAddress()
2020-01-31 11:23:33 +00:00
let value = c.stack.popInt()
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
result = (gas,
value,
2020-02-03 05:37:33 +00:00
destination,
c.msg.contractAddress, # sender
evmcCallCode,
2020-01-31 11:23:33 +00:00
c.stack.popInt().cleanMemRef,
c.stack.popInt().cleanMemRef,
c.stack.popInt().cleanMemRef,
c.stack.popInt().cleanMemRef,
c.msg.flags)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2020-02-03 05:37:33 +00:00
proc delegateCallParams(c: Computation): (UInt256, UInt256, EthAddress, EthAddress, CallKind, int, int, int, int, MsgFlags) =
let gas = c.stack.popInt()
2020-02-03 05:37:33 +00:00
let destination = c.stack.popAddress()
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
result = (gas,
c.msg.value, # value
2020-02-03 05:37:33 +00:00
destination,
c.msg.sender, # sender
evmcDelegateCall,
2020-01-31 11:23:33 +00:00
c.stack.popInt().cleanMemRef,
c.stack.popInt().cleanMemRef,
c.stack.popInt().cleanMemRef,
c.stack.popInt().cleanMemRef,
c.msg.flags)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2020-02-03 05:37:33 +00:00
proc staticCallParams(c: Computation): (UInt256, UInt256, EthAddress, EthAddress, CallKind, int, int, int, int, MsgFlags) =
let gas = c.stack.popInt()
2020-02-03 05:37:33 +00:00
let destination = c.stack.popAddress()
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
result = (gas,
0.u256, # value
2020-02-03 05:37:33 +00:00
destination,
c.msg.contractAddress, # sender
evmcCall,
2020-01-31 11:23:33 +00:00
c.stack.popInt().cleanMemRef,
c.stack.popInt().cleanMemRef,
c.stack.popInt().cleanMemRef,
c.stack.popInt().cleanMemRef,
emvcStatic) # is_static
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
template genCall(callName: untyped, opCode: Op): untyped =
2020-02-04 11:00:23 +00:00
op callName, inline = false:
## CALL, 0xf1, Message-Call into an account
## CALLCODE, 0xf2, Message-call into this account with an alternative account's code.
## DELEGATECALL, 0xf4, Message-call into this account with an alternative account's code, but persisting the current values for sender and value.
## STATICCALL, 0xfa, Static message-call into an account.
when opCode == Call:
2022-04-08 04:54:11 +00:00
if emvcStatic == c.msg.flags and c.stack[^3, UInt256] > 0.u256:
2020-02-04 11:00:23 +00:00
raise newException(StaticContextError, "Cannot modify state while inside of a STATICCALL context")
2020-02-03 05:37:33 +00:00
let (gas, value, destination, sender, callKind,
2020-02-04 11:00:23 +00:00
memInPos, memInLen, memOutPos, memOutLen, flags) = `callName Params`(c)
push: 0
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
2019-05-01 12:21:46 +00:00
let (memOffset, memLength) = if calcMemSize(memInPos, memInLen) > calcMemSize(memOutPos, memOutLen):
2019-02-06 07:46:09 +00:00
(memInPos, memInLen)
else:
(memOutPos, memOutLen)
2021-01-13 11:17:38 +00:00
# EIP2929
# This came before old gas calculator
# because it will affect `c.gasMeter.gasRemaining`
# and further `childGasLimit`
if c.fork >= FkBerlin:
# The WarmStorageReadCostEIP2929 (100) is already deducted in the form of a constant `gasCall`
let gasCost = c.accessGas(ColdAccountAccessCost - WarmStorageReadCost, 0,
destination)
if gasCost != 0:
c.gasMeter.consumeGas(gasCost, reason = "EIP2929 gasCall")
2021-01-13 11:17:38 +00:00
2020-02-03 05:37:33 +00:00
let contractAddress = when opCode in {Call, StaticCall}: destination else: c.msg.contractAddress
2020-12-11 10:51:17 +00:00
var (gasCost, childGasLimit) = c.gasCosts[opCode].c_handler(
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
value,
2019-03-18 10:24:25 +00:00
GasParams(kind: opCode,
2020-01-17 11:48:14 +00:00
c_isNewAccount: not c.accountExists(contractAddress),
c_gasBalance: c.gasMeter.gasRemaining,
2019-04-15 04:10:40 +00:00
c_contractGas: gas,
c_currentMemSize: c.memory.len,
2019-02-06 07:46:09 +00:00
c_memOffset: memOffset,
2019-03-18 10:24:25 +00:00
c_memLength: memLength
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
))
2019-02-06 07:46:09 +00:00
2021-01-11 08:33:30 +00:00
# EIP 2046: temporary disabled
2020-11-19 06:56:22 +00:00
# reduce gas fee for precompiles
# from 700 to 40
2021-01-11 07:53:51 +00:00
#when opCode == StaticCall:
# if c.fork >= FkBerlin and destination.toInt <= MaxPrecompilesAddr:
2020-12-11 13:26:55 +00:00
# gasCost = gasCost - 660.GasInt
2020-11-19 06:56:22 +00:00
2020-12-11 10:51:17 +00:00
if gasCost >= 0:
c.gasMeter.consumeGas(gasCost, reason = $opCode)
2019-02-20 12:31:17 +00:00
c.returnData.setLen(0)
2020-01-15 01:51:41 +00:00
if c.msg.depth >= MaxCallDepth:
debug "Computation Failure", reason = "Stack too deep", maximumDepth = MaxCallDepth, depth = c.msg.depth
# return unused gas
c.gasMeter.returnGas(childGasLimit)
return
2020-12-11 10:51:17 +00:00
if gasCost < 0 and childGasLimit <= 0:
2020-02-04 11:00:23 +00:00
raise newException(OutOfGas, "Gas not enough to perform calculation (" & callName.astToStr & ")")
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
c.memory.extend(memInPos, memInLen)
c.memory.extend(memOutPos, memOutLen)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
when opCode in {CallCode, Call}:
2020-01-16 15:07:04 +00:00
let senderBalance = c.getBalance(sender)
if senderBalance < value:
Tracing: Remove some trace messages that occur a lot during sync Disable some trace messages which appeared a lot in the output and probably aren't so useful any more, when block processing is functioning well at high speed. Turning on the trace level globally is useful to get a feel for what's happening, but only if each category is kept to a reasonable amount. As well as overwhelming the output so that it's hard to see general activity, some of these messages happen so much they severely slow down processing. Ones called every time an EVM opcode uses some gas are particularly extreme. These messages have all been chosen as things which are probably not useful any more (the relevant functionality has been debugged and is tested plenty). These have been commented out rather than removed. It may be that turning trace topics on/off, or other selection, is a better longer term solution, but that will require better command line options and good defaults for sure. (I think higher levels `tracev` and `tracevv` levels (extra verbose) would be more useful for this sort of deep tracing on request.) For now, enabling `--log-level:TRACE` on the command line is quite useful as long as we keep each category reasonable, and this patch tries to keep that balance. - Don't show "has transactions" on virtually every block imported. - Don't show "Sender" and "txHash" lines on every transaction processed. - Don't show "GAS CONSUMPTION" on every opcode executed", this is way too much. - Don't show "GAS RETURNED" and "GAS REFUND" on each contract call. - Don't show "op: Stop" on every Stop opcode, which means every transaction. - Don't show "Insufficient funds" whenever a contract can't call another. - Don't show "ECRecover", "SHA256 precompile", "RIPEMD160", "Identity" or even "Call precompile" every time a precompile is called. These are very well tested now. - Don't show "executeOpcodes error" whenever a contract returns an error. (This is changed to `trace` too, it's a normal event that is well tested.) Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-07-22 13:35:41 +00:00
#debug "Insufficient funds", available = senderBalance, needed = c.msg.value
# return unused gas
c.gasMeter.returnGas(childGasLimit)
return
2020-02-04 11:36:35 +00:00
when evmc_enabled:
EVMC: Small stacks when using EVMC, closes #575 (segfaults) This patch reduces stack space used with EVM in ENABLE_EVMC=1 mode, from 13 MB worst case to 550 kB, a 24x reduction. This completes fixing the "stack problem" and closes #575 (`EVM: Different segmentation faults when running the test suite with EVMC`). It also closes #256 (`recursive EVM call trigger unrecoverable stack overflow`). After this patch, it is possible to re-enable the CI targets which had to be disabled due to #575. This change is also a required precursor for switching over to "nearly EVMC" as the clean and focused Nimbus-internal API between EVM and sync/database processes, and is also key to the use of Chronos `async` in those processes when calling the EVM. (The motivation is the internal interface has to be substantially changed _anyway_ for the parallel sync and database processes, and EVMC turns out to be well-designed and well-suited for this. It provides good separation between modules, and suits our needs better than our other current interface. Might as well use a good one designed by someone else. EVMC is 98% done in Nimbus thanks to great work done before by @jangko, and we can use Nimbus-specific extensions where we need flexibility, including for performance. Being aligned with the ecosystem is a useful bonus feature.) All tests below were run on Ubuntu 20.04 LTS server, x86-64. This matches one of the targets that has been disabled for a while in CI in EVMC mode due to stack overflow crashing the tests, so it's a good choice. Measurements before =================== Testing commit `e76e0144 2021-04-22 11:29:42 +0700 add submodules: graphql and toml-serialization`. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 38416 depthHigh 3 ... Stack range 13074720 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 13.07 MB of stack to run, and so crash with the default stack limit on Ubuntu Server 20.04 (8MB). Exactly 12768 bytes per EVM call stack frame. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 14384 depthHigh 2 ... Stack range 3495456 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 3709600 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 7831600 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 7.83MB of stack to run. About 7648 bytes per EVM call stack frame. It _only just_ avoids crashing with the default Ubuntu Server stack limit of 8 MB. However, it still crashes on Windows x86-64, which is why the Windows CI EVMC target is currently disabled. On Linux where this passes, this is so borderline that it affects work and testing of the complex storage code, because that's called from the EVM. Also, this greatly exceeds the default thread stack size. Measurements after ================== $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 1936 depthHigh 3 ... Stack range 556272 depthHigh 1022 Stack range 556512 depthHigh 1023 Stack range 556816 depthHigh 1023 Stack range 557056 depthHigh 1024 Stack range 557360 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 1392 depthHigh 2 ... Stack range 248912 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 264144 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 557360 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stStaticCall/static_CallRecursiveBombPreCall.json For both tests, a satisfying *544 bytes* per EVM call stack frame, and EVM takes less than 600 kB total. With other overheads, both tests run in 600 kB stack total at maximum EVM depth. We must add some headroom on this for database activity called from the EVM, and different compile targets. But it means the EVM itself is no longer a stack burden. This is much smaller than the default thread stack size on Linux (2MB), with plenty of margin. (Just fyi, it isn't smaller than a _small_ thread stack on Linux from a long time ago (128kB), and some small embedded C targets.) This size is well suited to running EVMs in threads. Further reduction ================= This patch solves the stack problem. Windows and Linux 64-bit EVMC CI targets can be re-enabled, and there is no longer a problem with stack usage. We can reduce further to ~340 bytes per frame and 350 kB total, while still complying with EVMC. But as this involves changing how errors are handled to comply fully with EVMC, and removing `dispose` calls, it's not worth doing now while there are other EVMC changes in progress that will have the same effect. A Nimbus-specific extension will allow us to avoid recursion with EVMC anyway, bringing bytes per frame to zero. We need the extension anyway, to support Chronos `async` which parallel transaction processing is built around. Interop with non-Nimbus over EVMC won't let us avoid recursion, but then we can't control the stack frame size either. To prevent stack overflow in interop I anticipate using (this method in Aleth) [https://github.com/ethereum/aleth/blob/6e96ce34e3f131e2d42f3cb00741b54e05ab029d/libethereum/ExtVM.cpp#L61]. Smoke test other versions of GCC and Clang/LLVM =============================================== As all builds including Windows use GCC or Apple's Clang/LLVM, this is just to verify we're in the right ballpark on all targets. I've only checked `x86_64` though, not 32-bit, and not ARM. It's interesting to see GCC 10 uses less stack. This is because it optimises `struct` returns better, sometimes skipping an intermediate copy. Here it benefits the EVMC API, but I found GCC 10 also improves the larger stack usage of the rest of `nimbus-eth1` as well. Apple clang 12.0.0 (clang-1200.0.26.2) on MacOS 10.15: - 544 bytes per EVM call stack frame GCC 10.3.0 (Ubuntu 10.3.0-1ubuntu1) on Ubuntu 21.04: - 464 bytes per EVM call stack frame GCC 10.2.0 (Ubuntu 10.2.0-5ubuntu1~20.04) on Ubuntu 20.04 LTS: - 464 bytes per EVM call stack frame GCC 11.0.1 20210417 (experimental; Ubuntu 11-20210417-1ubuntu1) on Ubuntu 21.04: - 8 bytes per EVM call stack frame GCC 9.3.0 (Ubuntu 9.3.0-17ubuntu1~20.04) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 8.4.0 (Ubuntu 8.4.0-3ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 7.5.0 (Ubuntu 7.5.0-6ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 9.2.1 20191008 (Ubuntu 9.2.1-9ubuntu2) on Ubuntu 19.10: - 528 bytes per EVM call stack frame Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-04-26 15:01:00 +00:00
let msg = new(nimbus_message)
msg[] = nimbus_message(
2020-02-04 11:36:35 +00:00
kind: callKind.evmc_call_kind,
depth: (c.msg.depth + 1).int32,
gas: childGasLimit,
sender: sender,
destination: destination,
input_data: c.memory.readPtr(memInPos),
input_size: memInLen.uint,
value: toEvmc(value),
flags: flags.uint32
)
2020-02-03 05:37:33 +00:00
EVMC: Small stacks when using EVMC, closes #575 (segfaults) This patch reduces stack space used with EVM in ENABLE_EVMC=1 mode, from 13 MB worst case to 550 kB, a 24x reduction. This completes fixing the "stack problem" and closes #575 (`EVM: Different segmentation faults when running the test suite with EVMC`). It also closes #256 (`recursive EVM call trigger unrecoverable stack overflow`). After this patch, it is possible to re-enable the CI targets which had to be disabled due to #575. This change is also a required precursor for switching over to "nearly EVMC" as the clean and focused Nimbus-internal API between EVM and sync/database processes, and is also key to the use of Chronos `async` in those processes when calling the EVM. (The motivation is the internal interface has to be substantially changed _anyway_ for the parallel sync and database processes, and EVMC turns out to be well-designed and well-suited for this. It provides good separation between modules, and suits our needs better than our other current interface. Might as well use a good one designed by someone else. EVMC is 98% done in Nimbus thanks to great work done before by @jangko, and we can use Nimbus-specific extensions where we need flexibility, including for performance. Being aligned with the ecosystem is a useful bonus feature.) All tests below were run on Ubuntu 20.04 LTS server, x86-64. This matches one of the targets that has been disabled for a while in CI in EVMC mode due to stack overflow crashing the tests, so it's a good choice. Measurements before =================== Testing commit `e76e0144 2021-04-22 11:29:42 +0700 add submodules: graphql and toml-serialization`. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 38416 depthHigh 3 ... Stack range 13074720 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 13.07 MB of stack to run, and so crash with the default stack limit on Ubuntu Server 20.04 (8MB). Exactly 12768 bytes per EVM call stack frame. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 14384 depthHigh 2 ... Stack range 3495456 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 3709600 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 7831600 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 7.83MB of stack to run. About 7648 bytes per EVM call stack frame. It _only just_ avoids crashing with the default Ubuntu Server stack limit of 8 MB. However, it still crashes on Windows x86-64, which is why the Windows CI EVMC target is currently disabled. On Linux where this passes, this is so borderline that it affects work and testing of the complex storage code, because that's called from the EVM. Also, this greatly exceeds the default thread stack size. Measurements after ================== $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 1936 depthHigh 3 ... Stack range 556272 depthHigh 1022 Stack range 556512 depthHigh 1023 Stack range 556816 depthHigh 1023 Stack range 557056 depthHigh 1024 Stack range 557360 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 1392 depthHigh 2 ... Stack range 248912 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 264144 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 557360 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stStaticCall/static_CallRecursiveBombPreCall.json For both tests, a satisfying *544 bytes* per EVM call stack frame, and EVM takes less than 600 kB total. With other overheads, both tests run in 600 kB stack total at maximum EVM depth. We must add some headroom on this for database activity called from the EVM, and different compile targets. But it means the EVM itself is no longer a stack burden. This is much smaller than the default thread stack size on Linux (2MB), with plenty of margin. (Just fyi, it isn't smaller than a _small_ thread stack on Linux from a long time ago (128kB), and some small embedded C targets.) This size is well suited to running EVMs in threads. Further reduction ================= This patch solves the stack problem. Windows and Linux 64-bit EVMC CI targets can be re-enabled, and there is no longer a problem with stack usage. We can reduce further to ~340 bytes per frame and 350 kB total, while still complying with EVMC. But as this involves changing how errors are handled to comply fully with EVMC, and removing `dispose` calls, it's not worth doing now while there are other EVMC changes in progress that will have the same effect. A Nimbus-specific extension will allow us to avoid recursion with EVMC anyway, bringing bytes per frame to zero. We need the extension anyway, to support Chronos `async` which parallel transaction processing is built around. Interop with non-Nimbus over EVMC won't let us avoid recursion, but then we can't control the stack frame size either. To prevent stack overflow in interop I anticipate using (this method in Aleth) [https://github.com/ethereum/aleth/blob/6e96ce34e3f131e2d42f3cb00741b54e05ab029d/libethereum/ExtVM.cpp#L61]. Smoke test other versions of GCC and Clang/LLVM =============================================== As all builds including Windows use GCC or Apple's Clang/LLVM, this is just to verify we're in the right ballpark on all targets. I've only checked `x86_64` though, not 32-bit, and not ARM. It's interesting to see GCC 10 uses less stack. This is because it optimises `struct` returns better, sometimes skipping an intermediate copy. Here it benefits the EVMC API, but I found GCC 10 also improves the larger stack usage of the rest of `nimbus-eth1` as well. Apple clang 12.0.0 (clang-1200.0.26.2) on MacOS 10.15: - 544 bytes per EVM call stack frame GCC 10.3.0 (Ubuntu 10.3.0-1ubuntu1) on Ubuntu 21.04: - 464 bytes per EVM call stack frame GCC 10.2.0 (Ubuntu 10.2.0-5ubuntu1~20.04) on Ubuntu 20.04 LTS: - 464 bytes per EVM call stack frame GCC 11.0.1 20210417 (experimental; Ubuntu 11-20210417-1ubuntu1) on Ubuntu 21.04: - 8 bytes per EVM call stack frame GCC 9.3.0 (Ubuntu 9.3.0-17ubuntu1~20.04) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 8.4.0 (Ubuntu 8.4.0-3ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 7.5.0 (Ubuntu 7.5.0-6ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 9.2.1 20191008 (Ubuntu 9.2.1-9ubuntu2) on Ubuntu 19.10: - 528 bytes per EVM call stack frame Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-04-26 15:01:00 +00:00
c.chainTo(msg):
c.returnData = @(makeOpenArray(c.res.outputData, c.res.outputSize.int))
2020-02-03 05:37:33 +00:00
EVMC: Small stacks when using EVMC, closes #575 (segfaults) This patch reduces stack space used with EVM in ENABLE_EVMC=1 mode, from 13 MB worst case to 550 kB, a 24x reduction. This completes fixing the "stack problem" and closes #575 (`EVM: Different segmentation faults when running the test suite with EVMC`). It also closes #256 (`recursive EVM call trigger unrecoverable stack overflow`). After this patch, it is possible to re-enable the CI targets which had to be disabled due to #575. This change is also a required precursor for switching over to "nearly EVMC" as the clean and focused Nimbus-internal API between EVM and sync/database processes, and is also key to the use of Chronos `async` in those processes when calling the EVM. (The motivation is the internal interface has to be substantially changed _anyway_ for the parallel sync and database processes, and EVMC turns out to be well-designed and well-suited for this. It provides good separation between modules, and suits our needs better than our other current interface. Might as well use a good one designed by someone else. EVMC is 98% done in Nimbus thanks to great work done before by @jangko, and we can use Nimbus-specific extensions where we need flexibility, including for performance. Being aligned with the ecosystem is a useful bonus feature.) All tests below were run on Ubuntu 20.04 LTS server, x86-64. This matches one of the targets that has been disabled for a while in CI in EVMC mode due to stack overflow crashing the tests, so it's a good choice. Measurements before =================== Testing commit `e76e0144 2021-04-22 11:29:42 +0700 add submodules: graphql and toml-serialization`. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 38416 depthHigh 3 ... Stack range 13074720 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 13.07 MB of stack to run, and so crash with the default stack limit on Ubuntu Server 20.04 (8MB). Exactly 12768 bytes per EVM call stack frame. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 14384 depthHigh 2 ... Stack range 3495456 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 3709600 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 7831600 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 7.83MB of stack to run. About 7648 bytes per EVM call stack frame. It _only just_ avoids crashing with the default Ubuntu Server stack limit of 8 MB. However, it still crashes on Windows x86-64, which is why the Windows CI EVMC target is currently disabled. On Linux where this passes, this is so borderline that it affects work and testing of the complex storage code, because that's called from the EVM. Also, this greatly exceeds the default thread stack size. Measurements after ================== $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 1936 depthHigh 3 ... Stack range 556272 depthHigh 1022 Stack range 556512 depthHigh 1023 Stack range 556816 depthHigh 1023 Stack range 557056 depthHigh 1024 Stack range 557360 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 1392 depthHigh 2 ... Stack range 248912 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 264144 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 557360 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stStaticCall/static_CallRecursiveBombPreCall.json For both tests, a satisfying *544 bytes* per EVM call stack frame, and EVM takes less than 600 kB total. With other overheads, both tests run in 600 kB stack total at maximum EVM depth. We must add some headroom on this for database activity called from the EVM, and different compile targets. But it means the EVM itself is no longer a stack burden. This is much smaller than the default thread stack size on Linux (2MB), with plenty of margin. (Just fyi, it isn't smaller than a _small_ thread stack on Linux from a long time ago (128kB), and some small embedded C targets.) This size is well suited to running EVMs in threads. Further reduction ================= This patch solves the stack problem. Windows and Linux 64-bit EVMC CI targets can be re-enabled, and there is no longer a problem with stack usage. We can reduce further to ~340 bytes per frame and 350 kB total, while still complying with EVMC. But as this involves changing how errors are handled to comply fully with EVMC, and removing `dispose` calls, it's not worth doing now while there are other EVMC changes in progress that will have the same effect. A Nimbus-specific extension will allow us to avoid recursion with EVMC anyway, bringing bytes per frame to zero. We need the extension anyway, to support Chronos `async` which parallel transaction processing is built around. Interop with non-Nimbus over EVMC won't let us avoid recursion, but then we can't control the stack frame size either. To prevent stack overflow in interop I anticipate using (this method in Aleth) [https://github.com/ethereum/aleth/blob/6e96ce34e3f131e2d42f3cb00741b54e05ab029d/libethereum/ExtVM.cpp#L61]. Smoke test other versions of GCC and Clang/LLVM =============================================== As all builds including Windows use GCC or Apple's Clang/LLVM, this is just to verify we're in the right ballpark on all targets. I've only checked `x86_64` though, not 32-bit, and not ARM. It's interesting to see GCC 10 uses less stack. This is because it optimises `struct` returns better, sometimes skipping an intermediate copy. Here it benefits the EVMC API, but I found GCC 10 also improves the larger stack usage of the rest of `nimbus-eth1` as well. Apple clang 12.0.0 (clang-1200.0.26.2) on MacOS 10.15: - 544 bytes per EVM call stack frame GCC 10.3.0 (Ubuntu 10.3.0-1ubuntu1) on Ubuntu 21.04: - 464 bytes per EVM call stack frame GCC 10.2.0 (Ubuntu 10.2.0-5ubuntu1~20.04) on Ubuntu 20.04 LTS: - 464 bytes per EVM call stack frame GCC 11.0.1 20210417 (experimental; Ubuntu 11-20210417-1ubuntu1) on Ubuntu 21.04: - 8 bytes per EVM call stack frame GCC 9.3.0 (Ubuntu 9.3.0-17ubuntu1~20.04) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 8.4.0 (Ubuntu 8.4.0-3ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 7.5.0 (Ubuntu 7.5.0-6ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 9.2.1 20191008 (Ubuntu 9.2.1-9ubuntu2) on Ubuntu 19.10: - 528 bytes per EVM call stack frame Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-04-26 15:01:00 +00:00
let actualOutputSize = min(memOutLen, c.returnData.len)
if actualOutputSize > 0:
c.memory.write(memOutPos,
c.returnData.toOpenArray(0, actualOutputSize - 1))
2020-02-03 05:37:33 +00:00
EVMC: Small stacks when using EVMC, closes #575 (segfaults) This patch reduces stack space used with EVM in ENABLE_EVMC=1 mode, from 13 MB worst case to 550 kB, a 24x reduction. This completes fixing the "stack problem" and closes #575 (`EVM: Different segmentation faults when running the test suite with EVMC`). It also closes #256 (`recursive EVM call trigger unrecoverable stack overflow`). After this patch, it is possible to re-enable the CI targets which had to be disabled due to #575. This change is also a required precursor for switching over to "nearly EVMC" as the clean and focused Nimbus-internal API between EVM and sync/database processes, and is also key to the use of Chronos `async` in those processes when calling the EVM. (The motivation is the internal interface has to be substantially changed _anyway_ for the parallel sync and database processes, and EVMC turns out to be well-designed and well-suited for this. It provides good separation between modules, and suits our needs better than our other current interface. Might as well use a good one designed by someone else. EVMC is 98% done in Nimbus thanks to great work done before by @jangko, and we can use Nimbus-specific extensions where we need flexibility, including for performance. Being aligned with the ecosystem is a useful bonus feature.) All tests below were run on Ubuntu 20.04 LTS server, x86-64. This matches one of the targets that has been disabled for a while in CI in EVMC mode due to stack overflow crashing the tests, so it's a good choice. Measurements before =================== Testing commit `e76e0144 2021-04-22 11:29:42 +0700 add submodules: graphql and toml-serialization`. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 38416 depthHigh 3 ... Stack range 13074720 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 13.07 MB of stack to run, and so crash with the default stack limit on Ubuntu Server 20.04 (8MB). Exactly 12768 bytes per EVM call stack frame. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 14384 depthHigh 2 ... Stack range 3495456 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 3709600 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 7831600 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 7.83MB of stack to run. About 7648 bytes per EVM call stack frame. It _only just_ avoids crashing with the default Ubuntu Server stack limit of 8 MB. However, it still crashes on Windows x86-64, which is why the Windows CI EVMC target is currently disabled. On Linux where this passes, this is so borderline that it affects work and testing of the complex storage code, because that's called from the EVM. Also, this greatly exceeds the default thread stack size. Measurements after ================== $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 1936 depthHigh 3 ... Stack range 556272 depthHigh 1022 Stack range 556512 depthHigh 1023 Stack range 556816 depthHigh 1023 Stack range 557056 depthHigh 1024 Stack range 557360 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 1392 depthHigh 2 ... Stack range 248912 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 264144 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 557360 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stStaticCall/static_CallRecursiveBombPreCall.json For both tests, a satisfying *544 bytes* per EVM call stack frame, and EVM takes less than 600 kB total. With other overheads, both tests run in 600 kB stack total at maximum EVM depth. We must add some headroom on this for database activity called from the EVM, and different compile targets. But it means the EVM itself is no longer a stack burden. This is much smaller than the default thread stack size on Linux (2MB), with plenty of margin. (Just fyi, it isn't smaller than a _small_ thread stack on Linux from a long time ago (128kB), and some small embedded C targets.) This size is well suited to running EVMs in threads. Further reduction ================= This patch solves the stack problem. Windows and Linux 64-bit EVMC CI targets can be re-enabled, and there is no longer a problem with stack usage. We can reduce further to ~340 bytes per frame and 350 kB total, while still complying with EVMC. But as this involves changing how errors are handled to comply fully with EVMC, and removing `dispose` calls, it's not worth doing now while there are other EVMC changes in progress that will have the same effect. A Nimbus-specific extension will allow us to avoid recursion with EVMC anyway, bringing bytes per frame to zero. We need the extension anyway, to support Chronos `async` which parallel transaction processing is built around. Interop with non-Nimbus over EVMC won't let us avoid recursion, but then we can't control the stack frame size either. To prevent stack overflow in interop I anticipate using (this method in Aleth) [https://github.com/ethereum/aleth/blob/6e96ce34e3f131e2d42f3cb00741b54e05ab029d/libethereum/ExtVM.cpp#L61]. Smoke test other versions of GCC and Clang/LLVM =============================================== As all builds including Windows use GCC or Apple's Clang/LLVM, this is just to verify we're in the right ballpark on all targets. I've only checked `x86_64` though, not 32-bit, and not ARM. It's interesting to see GCC 10 uses less stack. This is because it optimises `struct` returns better, sometimes skipping an intermediate copy. Here it benefits the EVMC API, but I found GCC 10 also improves the larger stack usage of the rest of `nimbus-eth1` as well. Apple clang 12.0.0 (clang-1200.0.26.2) on MacOS 10.15: - 544 bytes per EVM call stack frame GCC 10.3.0 (Ubuntu 10.3.0-1ubuntu1) on Ubuntu 21.04: - 464 bytes per EVM call stack frame GCC 10.2.0 (Ubuntu 10.2.0-5ubuntu1~20.04) on Ubuntu 20.04 LTS: - 464 bytes per EVM call stack frame GCC 11.0.1 20210417 (experimental; Ubuntu 11-20210417-1ubuntu1) on Ubuntu 21.04: - 8 bytes per EVM call stack frame GCC 9.3.0 (Ubuntu 9.3.0-17ubuntu1~20.04) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 8.4.0 (Ubuntu 8.4.0-3ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 7.5.0 (Ubuntu 7.5.0-6ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 9.2.1 20191008 (Ubuntu 9.2.1-9ubuntu2) on Ubuntu 19.10: - 528 bytes per EVM call stack frame Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-04-26 15:01:00 +00:00
c.gasMeter.returnGas(c.res.gas_left)
2020-02-03 05:37:33 +00:00
EVMC: Small stacks when using EVMC, closes #575 (segfaults) This patch reduces stack space used with EVM in ENABLE_EVMC=1 mode, from 13 MB worst case to 550 kB, a 24x reduction. This completes fixing the "stack problem" and closes #575 (`EVM: Different segmentation faults when running the test suite with EVMC`). It also closes #256 (`recursive EVM call trigger unrecoverable stack overflow`). After this patch, it is possible to re-enable the CI targets which had to be disabled due to #575. This change is also a required precursor for switching over to "nearly EVMC" as the clean and focused Nimbus-internal API between EVM and sync/database processes, and is also key to the use of Chronos `async` in those processes when calling the EVM. (The motivation is the internal interface has to be substantially changed _anyway_ for the parallel sync and database processes, and EVMC turns out to be well-designed and well-suited for this. It provides good separation between modules, and suits our needs better than our other current interface. Might as well use a good one designed by someone else. EVMC is 98% done in Nimbus thanks to great work done before by @jangko, and we can use Nimbus-specific extensions where we need flexibility, including for performance. Being aligned with the ecosystem is a useful bonus feature.) All tests below were run on Ubuntu 20.04 LTS server, x86-64. This matches one of the targets that has been disabled for a while in CI in EVMC mode due to stack overflow crashing the tests, so it's a good choice. Measurements before =================== Testing commit `e76e0144 2021-04-22 11:29:42 +0700 add submodules: graphql and toml-serialization`. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 38416 depthHigh 3 ... Stack range 13074720 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 13.07 MB of stack to run, and so crash with the default stack limit on Ubuntu Server 20.04 (8MB). Exactly 12768 bytes per EVM call stack frame. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 14384 depthHigh 2 ... Stack range 3495456 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 3709600 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 7831600 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 7.83MB of stack to run. About 7648 bytes per EVM call stack frame. It _only just_ avoids crashing with the default Ubuntu Server stack limit of 8 MB. However, it still crashes on Windows x86-64, which is why the Windows CI EVMC target is currently disabled. On Linux where this passes, this is so borderline that it affects work and testing of the complex storage code, because that's called from the EVM. Also, this greatly exceeds the default thread stack size. Measurements after ================== $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 1936 depthHigh 3 ... Stack range 556272 depthHigh 1022 Stack range 556512 depthHigh 1023 Stack range 556816 depthHigh 1023 Stack range 557056 depthHigh 1024 Stack range 557360 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 1392 depthHigh 2 ... Stack range 248912 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 264144 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 557360 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stStaticCall/static_CallRecursiveBombPreCall.json For both tests, a satisfying *544 bytes* per EVM call stack frame, and EVM takes less than 600 kB total. With other overheads, both tests run in 600 kB stack total at maximum EVM depth. We must add some headroom on this for database activity called from the EVM, and different compile targets. But it means the EVM itself is no longer a stack burden. This is much smaller than the default thread stack size on Linux (2MB), with plenty of margin. (Just fyi, it isn't smaller than a _small_ thread stack on Linux from a long time ago (128kB), and some small embedded C targets.) This size is well suited to running EVMs in threads. Further reduction ================= This patch solves the stack problem. Windows and Linux 64-bit EVMC CI targets can be re-enabled, and there is no longer a problem with stack usage. We can reduce further to ~340 bytes per frame and 350 kB total, while still complying with EVMC. But as this involves changing how errors are handled to comply fully with EVMC, and removing `dispose` calls, it's not worth doing now while there are other EVMC changes in progress that will have the same effect. A Nimbus-specific extension will allow us to avoid recursion with EVMC anyway, bringing bytes per frame to zero. We need the extension anyway, to support Chronos `async` which parallel transaction processing is built around. Interop with non-Nimbus over EVMC won't let us avoid recursion, but then we can't control the stack frame size either. To prevent stack overflow in interop I anticipate using (this method in Aleth) [https://github.com/ethereum/aleth/blob/6e96ce34e3f131e2d42f3cb00741b54e05ab029d/libethereum/ExtVM.cpp#L61]. Smoke test other versions of GCC and Clang/LLVM =============================================== As all builds including Windows use GCC or Apple's Clang/LLVM, this is just to verify we're in the right ballpark on all targets. I've only checked `x86_64` though, not 32-bit, and not ARM. It's interesting to see GCC 10 uses less stack. This is because it optimises `struct` returns better, sometimes skipping an intermediate copy. Here it benefits the EVMC API, but I found GCC 10 also improves the larger stack usage of the rest of `nimbus-eth1` as well. Apple clang 12.0.0 (clang-1200.0.26.2) on MacOS 10.15: - 544 bytes per EVM call stack frame GCC 10.3.0 (Ubuntu 10.3.0-1ubuntu1) on Ubuntu 21.04: - 464 bytes per EVM call stack frame GCC 10.2.0 (Ubuntu 10.2.0-5ubuntu1~20.04) on Ubuntu 20.04 LTS: - 464 bytes per EVM call stack frame GCC 11.0.1 20210417 (experimental; Ubuntu 11-20210417-1ubuntu1) on Ubuntu 21.04: - 8 bytes per EVM call stack frame GCC 9.3.0 (Ubuntu 9.3.0-17ubuntu1~20.04) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 8.4.0 (Ubuntu 8.4.0-3ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 7.5.0 (Ubuntu 7.5.0-6ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 9.2.1 20191008 (Ubuntu 9.2.1-9ubuntu2) on Ubuntu 19.10: - 528 bytes per EVM call stack frame Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-04-26 15:01:00 +00:00
if c.res.status_code == EVMC_SUCCESS:
c.stack.top(1)
2020-02-03 05:37:33 +00:00
EVMC: Small stacks when using EVMC, closes #575 (segfaults) This patch reduces stack space used with EVM in ENABLE_EVMC=1 mode, from 13 MB worst case to 550 kB, a 24x reduction. This completes fixing the "stack problem" and closes #575 (`EVM: Different segmentation faults when running the test suite with EVMC`). It also closes #256 (`recursive EVM call trigger unrecoverable stack overflow`). After this patch, it is possible to re-enable the CI targets which had to be disabled due to #575. This change is also a required precursor for switching over to "nearly EVMC" as the clean and focused Nimbus-internal API between EVM and sync/database processes, and is also key to the use of Chronos `async` in those processes when calling the EVM. (The motivation is the internal interface has to be substantially changed _anyway_ for the parallel sync and database processes, and EVMC turns out to be well-designed and well-suited for this. It provides good separation between modules, and suits our needs better than our other current interface. Might as well use a good one designed by someone else. EVMC is 98% done in Nimbus thanks to great work done before by @jangko, and we can use Nimbus-specific extensions where we need flexibility, including for performance. Being aligned with the ecosystem is a useful bonus feature.) All tests below were run on Ubuntu 20.04 LTS server, x86-64. This matches one of the targets that has been disabled for a while in CI in EVMC mode due to stack overflow crashing the tests, so it's a good choice. Measurements before =================== Testing commit `e76e0144 2021-04-22 11:29:42 +0700 add submodules: graphql and toml-serialization`. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 38416 depthHigh 3 ... Stack range 13074720 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 13.07 MB of stack to run, and so crash with the default stack limit on Ubuntu Server 20.04 (8MB). Exactly 12768 bytes per EVM call stack frame. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 14384 depthHigh 2 ... Stack range 3495456 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 3709600 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 7831600 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 7.83MB of stack to run. About 7648 bytes per EVM call stack frame. It _only just_ avoids crashing with the default Ubuntu Server stack limit of 8 MB. However, it still crashes on Windows x86-64, which is why the Windows CI EVMC target is currently disabled. On Linux where this passes, this is so borderline that it affects work and testing of the complex storage code, because that's called from the EVM. Also, this greatly exceeds the default thread stack size. Measurements after ================== $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 1936 depthHigh 3 ... Stack range 556272 depthHigh 1022 Stack range 556512 depthHigh 1023 Stack range 556816 depthHigh 1023 Stack range 557056 depthHigh 1024 Stack range 557360 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 600 # Because we can! 600k stack. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 1392 depthHigh 2 ... Stack range 248912 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 264144 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 557360 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stStaticCall/static_CallRecursiveBombPreCall.json For both tests, a satisfying *544 bytes* per EVM call stack frame, and EVM takes less than 600 kB total. With other overheads, both tests run in 600 kB stack total at maximum EVM depth. We must add some headroom on this for database activity called from the EVM, and different compile targets. But it means the EVM itself is no longer a stack burden. This is much smaller than the default thread stack size on Linux (2MB), with plenty of margin. (Just fyi, it isn't smaller than a _small_ thread stack on Linux from a long time ago (128kB), and some small embedded C targets.) This size is well suited to running EVMs in threads. Further reduction ================= This patch solves the stack problem. Windows and Linux 64-bit EVMC CI targets can be re-enabled, and there is no longer a problem with stack usage. We can reduce further to ~340 bytes per frame and 350 kB total, while still complying with EVMC. But as this involves changing how errors are handled to comply fully with EVMC, and removing `dispose` calls, it's not worth doing now while there are other EVMC changes in progress that will have the same effect. A Nimbus-specific extension will allow us to avoid recursion with EVMC anyway, bringing bytes per frame to zero. We need the extension anyway, to support Chronos `async` which parallel transaction processing is built around. Interop with non-Nimbus over EVMC won't let us avoid recursion, but then we can't control the stack frame size either. To prevent stack overflow in interop I anticipate using (this method in Aleth) [https://github.com/ethereum/aleth/blob/6e96ce34e3f131e2d42f3cb00741b54e05ab029d/libethereum/ExtVM.cpp#L61]. Smoke test other versions of GCC and Clang/LLVM =============================================== As all builds including Windows use GCC or Apple's Clang/LLVM, this is just to verify we're in the right ballpark on all targets. I've only checked `x86_64` though, not 32-bit, and not ARM. It's interesting to see GCC 10 uses less stack. This is because it optimises `struct` returns better, sometimes skipping an intermediate copy. Here it benefits the EVMC API, but I found GCC 10 also improves the larger stack usage of the rest of `nimbus-eth1` as well. Apple clang 12.0.0 (clang-1200.0.26.2) on MacOS 10.15: - 544 bytes per EVM call stack frame GCC 10.3.0 (Ubuntu 10.3.0-1ubuntu1) on Ubuntu 21.04: - 464 bytes per EVM call stack frame GCC 10.2.0 (Ubuntu 10.2.0-5ubuntu1~20.04) on Ubuntu 20.04 LTS: - 464 bytes per EVM call stack frame GCC 11.0.1 20210417 (experimental; Ubuntu 11-20210417-1ubuntu1) on Ubuntu 21.04: - 8 bytes per EVM call stack frame GCC 9.3.0 (Ubuntu 9.3.0-17ubuntu1~20.04) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 8.4.0 (Ubuntu 8.4.0-3ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 7.5.0 (Ubuntu 7.5.0-6ubuntu2) on Ubuntu 20.04 LTS: - 544 bytes per EVM call stack frame GCC 9.2.1 20191008 (Ubuntu 9.2.1-9ubuntu2) on Ubuntu 19.10: - 528 bytes per EVM call stack frame Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-04-26 15:01:00 +00:00
if not c.res.release.isNil:
c.res.release(c.res)
2020-02-04 11:36:35 +00:00
else:
let msg = Message(
kind: callKind,
depth: c.msg.depth + 1,
gas: childGasLimit,
sender: sender,
contractAddress: contractAddress,
codeAddress: destination,
value: value,
data: c.memory.read(memInPos, memInLen),
flags: flags)
2020-02-03 05:37:33 +00:00
2020-02-04 11:36:35 +00:00
var child = newComputation(c.vmState, msg)
EVM: Small patch that reduces EVM stack usage to almost nothing There's been a lot of talk about the Nimbus EVM "stack problem". I think we assumed changing it would require big changes to the interpreter code, touching a lot of functions. It turned out to be a low hanging fruit. This patch solves the stack problem, but hardly touches anything. The change in EVM stack memory is from 13 MB worst case to just 48 kB, a 250x reduction. I've been doing work on the database/storage/trie code. While looking at the API between the EVM and the database/storage/trie, this stack patch stood out and made itself obvious. As it's tiny, rather than more talk, here it is. Note: This patch is intentionally small, non-invasive, and hopefully easy to understand, so that it doesn't conflict with other work done on the EVM, and can easily be grafted into any other EVM structure. Motivation ========== - We run out of space and crash on some targets, unless the stack limit is raised above its default. Surprise segmentation faults are unhelpful. - Some CI targets have been disabled for months due to this. - Because usage borders on the system limits, when working on database/storage/trie/sync code (called from the EVM), segmentation faults occur and are misleading. They cause lost time due to thinking there's a crash bug in the code being worked on, when there's nothing wrong with it. - Sometimes unrelated, trivial code changes elsewhere trigger CI test failures. It looks like abrupt termination. A simple, recent patch was crashing in `make test` even though it was a trivial refactor. Turns out it pushed the stack over the edge. - A large stack has to be scanned by the Nim garbage collector sometimes. Larger stack means slower GC and memory allocation. - The structure of this small patch suggests how to weave async into the EVM with almost no changes to the EVM, and no async transformation overhead. - The patch seemed obvious when working on the API between EVM and storage. Measurements before =================== All these tests were run on Ubuntu 20.04 server, x86-64. This is one of the targets that has been disabled for a while in CI in EVMC mode due to crashing, and excessive stack usage is the cause. Testing commit 0c34a8e3 `2021-04-08 17:46:00 +0200 CI: use MSYS2 on Windows`. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 38496 depthHigh 3 ... Stack range 13140272 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 13.14 MB of stack to run, and so crash with the default stack limit on Ubuntu Server 20.04 (8MB). Exactly 12832 bytes per EVM call stack frame. It's interesting to see some stack frames take a bit more. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 15488 depthHigh 2 ... Stack range 3539312 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 3756144 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 7929968 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 7.92MB of stack to run. About 7264 bytes per EVM call stack frame. It _only just_ avoids crashing with the default Ubuntu Server stack limit of 8 MB. However, it still crashes on Windows x86-64, which is why the CI target is currently disabled. On Linux where this passes, this is so borderline that it affects work and testing of storage and sync code, because that's called from the EVM. Which was a motivation for dealing with the stack instead of letting this linger. Also, this stack greatly exceeds the default thread stack size. $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 33216 depthHigh 3 ... Stack range 11338032 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 11.33 MB stack to run, and so crash with a default stack limit of 8MB. Exactly 11072 bytes per EVM call stack frame. It's interesting to see some stack frames take a bit more. $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 10224 depthHigh 2 ... Stack range 2471760 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 2623184 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 5537824 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 5.54 MB of stack to run, and avoid crashing on with a default stack limit of 8 MB. About 5408 bytes per EVM call stack frame. However, this is uncomfortably close to the limit, as the stack frame size is sensitive to changes in the code. Also, this stack greatly exceeds the default thread stack size. Measurements after ================== (This patch doesn't address EVMC mode, which is not our default. EVMC stack usage remains about the same. EVMC mode is addressed in another tiny patch.) $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 80 # Because we can! 80k stack. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 496 depthHigh 3 ... Stack range 49504 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 72 # Because we can! 72k stack. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 448 depthHigh 2 ... Stack range 22288 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 23632 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 49504 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json For both tests, a satisfying *48 bytes* per EVM call stack frame, and EVM takes not much more than 48 kB. With other overheads, both tests run in 80 kB stack total at maximum EVM depth. We must add some headroom on this for database activity called from the EVM, and different compile targets. But it means the EVM itself is no longer a stack burden. This is much smaller than the default thread stack size on Linux (2MB), with plenty of margin. It's even smaller than Linux from a long time ago (128kB), and some small embedded C targets. (Just fyi, though, some JVM environments allocated just 32 kB to thread stacks.) This size is also well suited to running EVMs in threads, if that's useful. Subtle exception handling and `dispose` ======================================= It is important that each `snapshot` has a corresponding `dispose` in the event of an exception being raised. This code does do that, but in a subtle way. The pair of functions `execCallOrCreate` and `execCallOrCreateAux` are equivalent to the following code, where you can see `dispose` more clearly: proc execCallOrCreate*(c: Computation) = defer: c.dispose() if c.beforeExec(): return c.executeOpcodes() while not c.continuation.isNil: c.child.execCallOrCreate() c.child = nil (c.continuation)() c.executeOpcodes() c.afterExec() That works fine, but only reduces the stack used to 300-700 kB instead of 48 kB. To get lower we split the above into separate `execCallOrCreate` and `execCallOrCreateAux`. Only the outermost has `defer`, and instead of handling one level, it walks the entire `c.parent` chain calling `dispose` if needed. The inner one avoids `defer`, which greatly reduces the size of its stackframe. `c` is a `var` parameter, at each level of recursion. So the outermost proc sees the temporary changes made by all inner calls. This is why `c` is updated and the `c.parent` chain is maintained at each step. Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-04-12 17:06:31 +00:00
c.chainTo(child):
if not child.shouldBurnGas:
c.gasMeter.returnGas(child.gasMeter.gasRemaining)
2020-02-03 05:37:33 +00:00
EVM: Small patch that reduces EVM stack usage to almost nothing There's been a lot of talk about the Nimbus EVM "stack problem". I think we assumed changing it would require big changes to the interpreter code, touching a lot of functions. It turned out to be a low hanging fruit. This patch solves the stack problem, but hardly touches anything. The change in EVM stack memory is from 13 MB worst case to just 48 kB, a 250x reduction. I've been doing work on the database/storage/trie code. While looking at the API between the EVM and the database/storage/trie, this stack patch stood out and made itself obvious. As it's tiny, rather than more talk, here it is. Note: This patch is intentionally small, non-invasive, and hopefully easy to understand, so that it doesn't conflict with other work done on the EVM, and can easily be grafted into any other EVM structure. Motivation ========== - We run out of space and crash on some targets, unless the stack limit is raised above its default. Surprise segmentation faults are unhelpful. - Some CI targets have been disabled for months due to this. - Because usage borders on the system limits, when working on database/storage/trie/sync code (called from the EVM), segmentation faults occur and are misleading. They cause lost time due to thinking there's a crash bug in the code being worked on, when there's nothing wrong with it. - Sometimes unrelated, trivial code changes elsewhere trigger CI test failures. It looks like abrupt termination. A simple, recent patch was crashing in `make test` even though it was a trivial refactor. Turns out it pushed the stack over the edge. - A large stack has to be scanned by the Nim garbage collector sometimes. Larger stack means slower GC and memory allocation. - The structure of this small patch suggests how to weave async into the EVM with almost no changes to the EVM, and no async transformation overhead. - The patch seemed obvious when working on the API between EVM and storage. Measurements before =================== All these tests were run on Ubuntu 20.04 server, x86-64. This is one of the targets that has been disabled for a while in CI in EVMC mode due to crashing, and excessive stack usage is the cause. Testing commit 0c34a8e3 `2021-04-08 17:46:00 +0200 CI: use MSYS2 on Windows`. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 38496 depthHigh 3 ... Stack range 13140272 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 13.14 MB of stack to run, and so crash with the default stack limit on Ubuntu Server 20.04 (8MB). Exactly 12832 bytes per EVM call stack frame. It's interesting to see some stack frames take a bit more. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 15488 depthHigh 2 ... Stack range 3539312 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 3756144 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 7929968 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 7.92MB of stack to run. About 7264 bytes per EVM call stack frame. It _only just_ avoids crashing with the default Ubuntu Server stack limit of 8 MB. However, it still crashes on Windows x86-64, which is why the CI target is currently disabled. On Linux where this passes, this is so borderline that it affects work and testing of storage and sync code, because that's called from the EVM. Which was a motivation for dealing with the stack instead of letting this linger. Also, this stack greatly exceeds the default thread stack size. $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 33216 depthHigh 3 ... Stack range 11338032 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 11.33 MB stack to run, and so crash with a default stack limit of 8MB. Exactly 11072 bytes per EVM call stack frame. It's interesting to see some stack frames take a bit more. $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 10224 depthHigh 2 ... Stack range 2471760 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 2623184 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 5537824 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 5.54 MB of stack to run, and avoid crashing on with a default stack limit of 8 MB. About 5408 bytes per EVM call stack frame. However, this is uncomfortably close to the limit, as the stack frame size is sensitive to changes in the code. Also, this stack greatly exceeds the default thread stack size. Measurements after ================== (This patch doesn't address EVMC mode, which is not our default. EVMC stack usage remains about the same. EVMC mode is addressed in another tiny patch.) $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 80 # Because we can! 80k stack. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 496 depthHigh 3 ... Stack range 49504 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 72 # Because we can! 72k stack. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 448 depthHigh 2 ... Stack range 22288 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 23632 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 49504 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json For both tests, a satisfying *48 bytes* per EVM call stack frame, and EVM takes not much more than 48 kB. With other overheads, both tests run in 80 kB stack total at maximum EVM depth. We must add some headroom on this for database activity called from the EVM, and different compile targets. But it means the EVM itself is no longer a stack burden. This is much smaller than the default thread stack size on Linux (2MB), with plenty of margin. It's even smaller than Linux from a long time ago (128kB), and some small embedded C targets. (Just fyi, though, some JVM environments allocated just 32 kB to thread stacks.) This size is also well suited to running EVMs in threads, if that's useful. Subtle exception handling and `dispose` ======================================= It is important that each `snapshot` has a corresponding `dispose` in the event of an exception being raised. This code does do that, but in a subtle way. The pair of functions `execCallOrCreate` and `execCallOrCreateAux` are equivalent to the following code, where you can see `dispose` more clearly: proc execCallOrCreate*(c: Computation) = defer: c.dispose() if c.beforeExec(): return c.executeOpcodes() while not c.continuation.isNil: c.child.execCallOrCreate() c.child = nil (c.continuation)() c.executeOpcodes() c.afterExec() That works fine, but only reduces the stack used to 300-700 kB instead of 48 kB. To get lower we split the above into separate `execCallOrCreate` and `execCallOrCreateAux`. Only the outermost has `defer`, and instead of handling one level, it walks the entire `c.parent` chain calling `dispose` if needed. The inner one avoids `defer`, which greatly reduces the size of its stackframe. `c` is a `var` parameter, at each level of recursion. So the outermost proc sees the temporary changes made by all inner calls. This is why `c` is updated and the `c.parent` chain is maintained at each step. Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-04-12 17:06:31 +00:00
if child.isSuccess:
c.merge(child)
c.stack.top(1)
2020-02-04 11:36:35 +00:00
EVM: Small patch that reduces EVM stack usage to almost nothing There's been a lot of talk about the Nimbus EVM "stack problem". I think we assumed changing it would require big changes to the interpreter code, touching a lot of functions. It turned out to be a low hanging fruit. This patch solves the stack problem, but hardly touches anything. The change in EVM stack memory is from 13 MB worst case to just 48 kB, a 250x reduction. I've been doing work on the database/storage/trie code. While looking at the API between the EVM and the database/storage/trie, this stack patch stood out and made itself obvious. As it's tiny, rather than more talk, here it is. Note: This patch is intentionally small, non-invasive, and hopefully easy to understand, so that it doesn't conflict with other work done on the EVM, and can easily be grafted into any other EVM structure. Motivation ========== - We run out of space and crash on some targets, unless the stack limit is raised above its default. Surprise segmentation faults are unhelpful. - Some CI targets have been disabled for months due to this. - Because usage borders on the system limits, when working on database/storage/trie/sync code (called from the EVM), segmentation faults occur and are misleading. They cause lost time due to thinking there's a crash bug in the code being worked on, when there's nothing wrong with it. - Sometimes unrelated, trivial code changes elsewhere trigger CI test failures. It looks like abrupt termination. A simple, recent patch was crashing in `make test` even though it was a trivial refactor. Turns out it pushed the stack over the edge. - A large stack has to be scanned by the Nim garbage collector sometimes. Larger stack means slower GC and memory allocation. - The structure of this small patch suggests how to weave async into the EVM with almost no changes to the EVM, and no async transformation overhead. - The patch seemed obvious when working on the API between EVM and storage. Measurements before =================== All these tests were run on Ubuntu 20.04 server, x86-64. This is one of the targets that has been disabled for a while in CI in EVMC mode due to crashing, and excessive stack usage is the cause. Testing commit 0c34a8e3 `2021-04-08 17:46:00 +0200 CI: use MSYS2 on Windows`. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 38496 depthHigh 3 ... Stack range 13140272 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 13.14 MB of stack to run, and so crash with the default stack limit on Ubuntu Server 20.04 (8MB). Exactly 12832 bytes per EVM call stack frame. It's interesting to see some stack frames take a bit more. $ rm -f build/all_tests && make ENABLE_EVMC=1 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 15488 depthHigh 2 ... Stack range 3539312 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 3756144 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 7929968 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 7.92MB of stack to run. About 7264 bytes per EVM call stack frame. It _only just_ avoids crashing with the default Ubuntu Server stack limit of 8 MB. However, it still crashes on Windows x86-64, which is why the CI target is currently disabled. On Linux where this passes, this is so borderline that it affects work and testing of storage and sync code, because that's called from the EVM. Which was a motivation for dealing with the stack instead of letting this linger. Also, this stack greatly exceeds the default thread stack size. $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 16384 # Requires larger stack than default to avoid crash. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 33216 depthHigh 3 ... Stack range 11338032 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json These tests use 11.33 MB stack to run, and so crash with a default stack limit of 8MB. Exactly 11072 bytes per EVM call stack frame. It's interesting to see some stack frames take a bit more. $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 16384 # Requires larger stack than default. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 10224 depthHigh 2 ... Stack range 2471760 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 2623184 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 5537824 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json These tests use 5.54 MB of stack to run, and avoid crashing on with a default stack limit of 8 MB. About 5408 bytes per EVM call stack frame. However, this is uncomfortably close to the limit, as the stack frame size is sensitive to changes in the code. Also, this stack greatly exceeds the default thread stack size. Measurements after ================== (This patch doesn't address EVMC mode, which is not our default. EVMC stack usage remains about the same. EVMC mode is addressed in another tiny patch.) $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 80 # Because we can! 80k stack. $ ./build/all_tests 9 | tee tlog [Suite] persist block json tests ... Stack range 496 depthHigh 3 ... Stack range 49504 depthHigh 1024 [OK] tests/fixtures/PersistBlockTests/block1431916.json $ rm -f build/all_tests && make ENABLE_EVMC=0 test $ ulimit -S -s 72 # Because we can! 72k stack. $ ./build/all_tests 7 | tee tlog [Suite] new generalstate json tests ... Stack range 448 depthHigh 2 ... Stack range 22288 depthHigh 457 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest639.json ... Stack range 23632 depthHigh 485 [OK] tests/fixtures/eth_tests/GeneralStateTests/stRandom2/randomStatetest458.json ... Stack range 49504 depthHigh 1024 [OK] tests/fixtures/eth_tests/GeneralStateTests/stCreate2/Create2OnDepth1024.json For both tests, a satisfying *48 bytes* per EVM call stack frame, and EVM takes not much more than 48 kB. With other overheads, both tests run in 80 kB stack total at maximum EVM depth. We must add some headroom on this for database activity called from the EVM, and different compile targets. But it means the EVM itself is no longer a stack burden. This is much smaller than the default thread stack size on Linux (2MB), with plenty of margin. It's even smaller than Linux from a long time ago (128kB), and some small embedded C targets. (Just fyi, though, some JVM environments allocated just 32 kB to thread stacks.) This size is also well suited to running EVMs in threads, if that's useful. Subtle exception handling and `dispose` ======================================= It is important that each `snapshot` has a corresponding `dispose` in the event of an exception being raised. This code does do that, but in a subtle way. The pair of functions `execCallOrCreate` and `execCallOrCreateAux` are equivalent to the following code, where you can see `dispose` more clearly: proc execCallOrCreate*(c: Computation) = defer: c.dispose() if c.beforeExec(): return c.executeOpcodes() while not c.continuation.isNil: c.child.execCallOrCreate() c.child = nil (c.continuation)() c.executeOpcodes() c.afterExec() That works fine, but only reduces the stack used to 300-700 kB instead of 48 kB. To get lower we split the above into separate `execCallOrCreate` and `execCallOrCreateAux`. Only the outermost has `defer`, and instead of handling one level, it walks the entire `c.parent` chain calling `dispose` if needed. The inner one avoids `defer`, which greatly reduces the size of its stackframe. `c` is a `var` parameter, at each level of recursion. So the outermost proc sees the temporary changes made by all inner calls. This is why `c` is updated and the `c.parent` chain is maintained at each step. Signed-off-by: Jamie Lokier <jamie@shareable.org>
2021-04-12 17:06:31 +00:00
c.returnData = child.output
let actualOutputSize = min(memOutLen, child.output.len)
if actualOutputSize > 0:
c.memory.write(memOutPos,
child.output.toOpenArray(0, actualOutputSize - 1))
2020-02-04 11:36:35 +00:00
genCall(call, Call)
genCall(callCode, CallCode)
genCall(delegateCall, DelegateCall)
genCall(staticCall, StaticCall)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op returnOp, inline = false, startPos, size:
## 0xf3, Halt execution returning output data.
let (pos, len) = (startPos.cleanMemRef, size.cleanMemRef)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
c.gasMeter.consumeGas(
c.gasCosts[Return].m_handler(c.memory.len, pos, len),
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
reason = "RETURN"
)
c.memory.extend(pos, len)
c.output = c.memory.read(pos, len)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op revert, inline = false, startPos, size:
## 0xfd, Halt execution reverting state changes but returning data and remaining gas.
let (pos, len) = (startPos.cleanMemRef, size.cleanMemRef)
c.gasMeter.consumeGas(
c.gasCosts[Revert].m_handler(c.memory.len, pos, len),
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
reason = "REVERT"
)
c.memory.extend(pos, len)
c.output = c.memory.read(pos, len)
2019-04-22 09:32:50 +00:00
# setError(msg, false) will signal cheap revert
c.setError("REVERT opcode executed", false)
Refactor interpreter dispatch (#65) * move forks constants, rename errors * Move vm/utils to vm/interpreter/utils * initial opcodes refactoring * Add refactored Comparison & Bitwise Logic Operations * Add sha3 and address, simplify macro, support pop 0 * balance, origin, caller, callValue * fix gas copy opcodes gas costs, add callDataLoad/Size/Copy, CodeSize/Copy and gas price opcode * Update with 30s, 40s, 50s opcodes + impl of balance + stack improvement * add push, dup, swap, log, create and call operations * finish opcode implementation * Add the new dispatching logic * Pass the opcode test * Make test_vm_json compile * halt execution without exceptions for Return, Revert, selfdestruct (fix #62) * Properly catch and recover from EVM exceptions (stack underflow ...) * Fix byte op * Fix jump regressions * Update for latest devel, don't import old dispatch code as quasiBoolean macro is broken by latest devel * Fix sha3 regression on empty memory slice and until end of range slice * Fix padding / range error on expXY_success (gas computation left) * update logging procs * Add tracing - expXY_success is not a regression, sload stub was accidentally passing the test * Reuse the same stub as OO implementation * Delete previous opcode implementation * Delete object oriented fork code * Delete exceptions that were used as control flows * delete base.nim :fire:, yet another OO remnants * Delete opcode table * Enable omputed gotos and compile-time gas fees * Revert const gasCosts -> generates SIGSEGV * inline push, swap and dup opcodes * loggers are now template again, why does this pass new tests? * Trigger CI rebuild after rocksdb fix https://github.com/status-im/nim-rocksdb/pull/5 * Address review comment on "push" + VMTests in debug mode (not release) * Address review comment: don't tag fork by default, make opcode impl grepable * Static compilation fixes after rebasing * fix the initialization of the VM database * add a missing import * Deactivate balance and sload test following #59 * Reactivate stack check (deactivated in #59, necessary to pass tests) * Merge remaining opcodes implementation from #59 * Merge callDataLoad and codeCopy fixes, todo simplify see #67
2018-07-06 07:52:31 +00:00
op selfDestruct, inline = false:
2020-01-17 11:58:03 +00:00
## 0xff Halt execution and register account for later deletion.
let beneficiary = c.stack.popAddress()
2020-01-17 11:58:03 +00:00
c.selfDestruct(beneficiary)
op selfDestructEip150, inline = false:
let beneficiary = c.stack.popAddress()
let gasParams = GasParams(kind: SelfDestruct,
2020-01-16 14:43:29 +00:00
sd_condition: not c.accountExists(beneficiary)
)
let gasCost = c.gasCosts[SelfDestruct].c_handler(0.u256, gasParams).gasCost
c.gasMeter.consumeGas(gasCost, reason = "SELFDESTRUCT EIP150")
2020-01-17 11:58:03 +00:00
c.selfDestruct(beneficiary)
op selfDestructEip161, inline = false:
checkInStaticContext(c)
2019-04-23 13:52:53 +00:00
let
beneficiary = c.stack.popAddress()
2020-01-17 11:48:14 +00:00
isDead = not c.accountExists(beneficiary)
2020-01-16 15:07:04 +00:00
balance = c.getBalance(c.msg.contractAddress)
let gasParams = GasParams(kind: SelfDestruct,
sd_condition: isDead and not balance.isZero
)
let gasCost = c.gasCosts[SelfDestruct].c_handler(0.u256, gasParams).gasCost
c.gasMeter.consumeGas(gasCost, reason = "SELFDESTRUCT EIP161")
2020-01-17 11:58:03 +00:00
c.selfDestruct(beneficiary)
# Constantinople's new opcodes
op shlOp, inline = true, shift, num:
let shiftLen = shift.safeInt
if shiftLen >= 256:
push: 0
else:
push: num shl shiftLen
op shrOp, inline = true, shift, num:
let shiftLen = shift.safeInt
if shiftLen >= 256:
push: 0
else:
2019-08-12 11:22:22 +00:00
# uint version of `shr`
push: num shr shiftLen
op sarOp, inline = true:
let shiftLen = c.stack.popInt().safeInt
let num = cast[Int256](c.stack.popInt())
if shiftLen >= 256:
2019-05-11 14:44:06 +00:00
if num.isNegative:
2022-04-08 04:54:11 +00:00
push: cast[UInt256]((-1).i256)
else:
push: 0
else:
2019-08-12 11:22:22 +00:00
# int version of `shr` then force the result
# into uint256
2022-04-08 04:54:11 +00:00
push: cast[UInt256](num shr shiftLen)
op extCodeHash, inline = true:
let address = c.stack.popAddress()
2020-01-16 15:48:22 +00:00
push: c.getCodeHash(address)
2020-12-11 10:51:17 +00:00
op balanceEIP2929, inline = true:
## 0x31, Get balance of the given account.
let address = c.stack.popAddress()
c.gasMeter.consumeGas(c.accessGas(address),
reason = "balanceEIP2929")
2020-12-11 10:51:17 +00:00
push: c.getBalance(address)
op extCodeHashEIP2929, inline = true:
let address = c.stack.popAddress()
c.gasMeter.consumeGas(c.accessGas(address),
reason = "extCodeHashEIP2929")
2020-12-11 10:51:17 +00:00
push: c.getCodeHash(address)
op extCodeSizeEIP2929, inline = true:
## 0x3b, Get size of an account's code
let address = c.stack.popAddress()
c.gasMeter.consumeGas(c.accessGas(address),
reason = "extCodeSizeEIP2929")
2020-12-11 10:51:17 +00:00
push: c.getCodeSize(address)
op extCodeCopyEIP2929, inline = true:
## 0x3c, Copy an account's code to memory.
let address = c.stack.popAddress()
let (memStartPos, codeStartPos, size) = c.stack.popInt(3)
let (memPos, codePos, len) = (memStartPos.cleanMemRef, codeStartPos.cleanMemRef, size.cleanMemRef)
c.gasMeter.consumeGas(
c.gasCosts[ExtCodeCopy].m_handler(c.memory.len, memPos, len),
reason="ExtCodeCopy fee")
c.gasMeter.consumeGas(c.accessGas(address),
reason = "extCodeCopyEIP2929")
2020-12-11 10:51:17 +00:00
let codeBytes = c.getCode(address)
c.memory.writePaddedResult(codeBytes, memPos, codePos, len)
op selfDestructEIP2929, inline = false:
checkInStaticContext(c)
let
beneficiary = c.stack.popAddress()
isDead = not c.accountExists(beneficiary)
balance = c.getBalance(c.msg.contractAddress)
let gasParams = GasParams(kind: SelfDestruct,
sd_condition: isDead and not balance.isZero
)
var gasCost = c.gasCosts[SelfDestruct].c_handler(0.u256, gasParams).gasCost
gasCost += c.accessGas(ColdAccountAccessCost, 0, beneficiary)
2020-12-11 10:51:17 +00:00
c.gasMeter.consumeGas(gasCost, reason = "SELFDESTRUCT EIP161")
c.selfDestruct(beneficiary)
op sloadEIP2929, inline = true, slot:
## 0x54, Load word from storage.
let gasCost = c.accessGas(ColdSloadCost, WarmStorageReadCost, slot)
c.gasMeter.consumeGas(gasCost, reason = "sloadEIP2929")
2020-12-11 10:51:17 +00:00
push: c.getStorage(slot)
op sstoreEIP2929, inline = false, slot, newValue:
checkInStaticContext(c)
const SentryGasEIP2200 = 2300 # Minimum gas required to be present for an SSTORE call, not consumed
if c.gasMeter.gasRemaining <= SentryGasEIP2200:
raise newException(OutOfGas, "Gas not enough to perform EIP2200 SSTORE")
let gasCost = c.accessGas(ColdSloadCost, 0, slot)
if gasCost != 0:
c.gasMeter.consumeGas(gasCost, reason = "sstoreEIP2929")
2020-12-11 10:51:17 +00:00
when evmc_enabled:
sstoreEvmc(c, slot, newValue)
else:
sstoreNetGasMeteringImpl(c, slot, newValue)