EIP-3860: Limit and meter initcode

[yperbasis]

EIP-3860 is included into Shanghai. This PR implements Rules 3 & 4 of the specification. Rules 1 & 2 should be implemented on the host side. See erigontech/silkworm#808 for an example of host implementation.

[codecov]

Codecov Report

Merging #525 (0298d8a) into instr_create (4ea026b) will decrease coverage by 0.05%.
The diff coverage is 95.77%.

Additional details and impacted files

@@               Coverage Diff                @@
##           instr_create     #525      +/-   ##
================================================
- Coverage         97.45%   97.40%   -0.06%     
================================================
  Files                62       63       +1     
  Lines              5980     6047      +67     
================================================
+ Hits               5828     5890      +62     
- Misses              152      157       +5     

Flag	Coverage Δ
blockchaintests	76.80% <60.00%> (-0.12%)	⬇️
statetests	72.64% <75.00%> (-0.05%)	⬇️
unittests	89.82% <87.32%> (+0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
test/state/host.hpp	40.00% <ø> (ø)
test/state/state.cpp	77.64% <62.50%> (-2.10%)	⬇️
lib/evmone/instructions_calls.cpp	100.00% <100.00%> (ø)
test/state/host.cpp	80.54% <100.00%> (ø)
test/unittests/evm_eip3860_initcode_test.cpp	100.00% <100.00%> (ø)
test/utils/bytecode.hpp	96.37% <100.00%> (+0.03%)	⬆️
lib/evmone/advanced_analysis.hpp	97.05% <0.00%> (-2.95%)	⬇️
lib/evmone/instructions.hpp	99.79% <0.00%> (-0.21%)	⬇️

[chfast]

I'd add rule 3 here as well (before rule 4). I like to have "light" checks on the EVM side. This also helps against multiplication overflow. Any counter-arguments?

[yperbasis]

I'd add rule 3 here as well (before rule 4). I like to have "light" checks on the EVM side. This also helps against multiplication overflow. Any counter-arguments?

Initially I implemented rule 3 in evmone, but that deviates from geth's implementation ethereum/go-ethereum#23847 (see core/vm/evm.go). Specifically, that leads to an incorrect nonce in the createInitCodeSizeLimit_d1g0v0_Merge+3860 consensus test.

[yperbasis]

I've updated this PR according to ethereum/EIPs#6040.

The order of the checks in CREATE2 are going to change in the spec.

[chfast]

I messed up with GitHub here. Can you re-create the PR?

[yperbasis]

I messed up with GitHub here. Can you re-create the PR?

Sure. Re-created as PR #545.