fix: remove contradictory gas limit check

[mooori]

validate_chunk_with_chunk_extra_and_receipts_root compares prev_chunk_extra.gas_limit() and chunk_header.gas_limit() twice:

1. In the first comparison they must be equal.
2. A few lines down they might differ within the bounds of GAS_LIMIT_ADJUSTMENT_FACTOR.

If I'm not missing anything here, one of these checks should be removed. Since there is a GAS_LIMIT_ADJUSTMENT_FACTOR, I assume gas limit changes should be possible, so this PR removes the first check. Or has anything changed that requires gas limits of prev_chunk_extra and chunk_header to be equal now?

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 71.62%. Comparing base (566e926) to head (02fa4d2).

Additional details and impacted files

@@            Coverage Diff             @@
##           master   #11958      +/-   ##
==========================================
- Coverage   71.62%   71.62%   -0.01%     
==========================================
  Files         810      810              
  Lines      163751   163747       -4     
  Branches   163751   163747       -4     
==========================================
- Hits       117288   117283       -5     
- Misses      41387    41389       +2     
+ Partials     5076     5075       -1     

Flag	Coverage Δ
backward-compatibility	0.23% <ø> (+<0.01%)	⬆️
db-migration	0.23% <ø> (+<0.01%)	⬆️
genesis-check	1.33% <ø> (+<0.01%)	⬆️
integration-tests	38.53% <ø> (-0.01%)	⬇️
linux	71.39% <ø> (+0.01%)	⬆️
linux-nightly	71.19% <ø> (-0.02%)	⬇️
macos	53.14% <ø> (-1.20%)	⬇️
pytests	1.60% <ø> (+<0.01%)	⬆️
sanity-checks	1.40% <ø> (+<0.01%)	⬆️
unittests	65.69% <ø> (-0.01%)	⬇️
upgradability	0.28% <ø> (+<0.01%)	⬆️

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

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[wacban]

You are right that one of the checks is redundant.

I'm almost certain that the gas limit in prev_extra should be equal to the header gas limit. As a rule the chunk header contains info from the prev extra copied byte by byte.

I am not aware of any feature to adjust the gas limit. I know we have adjustments for gas price built in so perhaps there was a mismatch there? Can you do some digging and see if the gas limit can actually be changed and if the gas price adjustment is correctly verified?

[mooori]

My understanding is that gas_limit is passed through many layers, but so far it's never modified. Which would explain why these checks haven't caused any problems. For #11460 we're working on enabling gas limit adjustments in tests/benchmarks and #11863 is the first step towards that.

I'm almost certain that the gas limit in prev_extra should be equal to the header gas limit. As a rule the chunk header contains info from the prev extra copied byte by byte.

That's interesting...and leaves me confused as to why GAS_LIMIT_ADJUSTMENT_FACTOR exists? Because the only place where it is used is in the checks in this file.

Can you do some digging and see if the gas limit can actually be changed and if the gas price adjustment is correctly verified?

I've been able to change the gas limit. Doing it here triggers the following check to fail.

prev_chunk_extra.gas_used() != chunk_header.prev_gas_used()

When commenting out this check, nodes seem to run fine with gas limits that are adjusted repeatedly at runtime.

I'm taking the hint you gave in #11863 and will try to see if there is a place where the gas limit can be changed at runtime without triggering this test to fail.

[Longarithm]

LGTM due to #11863 (comment) but tagging @bowenwang1996 just in case.