precompiles: Use g1_mul and g2_mul for single-input multiplication

[pdobacz]

I'm new to the bls party, but the coverage was telling me these functions are uncovered, and the EIP told me we should dispatch to them: https://eips.ethereum.org/EIPS/eip-2537#no-dedicated-mul-call

So a quick take - is this the way to go? (bls.cpp is at 100% now)

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 87.31%. Comparing base (8b957b3) to head (49e209b).
⚠️ Report is 2 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1340      +/-   ##
==========================================
+ Coverage   87.27%   87.31%   +0.04%     
==========================================
  Files         169      169              
  Lines       25042    25058      +16     
  Branches     4118     4122       +4     
==========================================
+ Hits        21855    21879      +24     
+ Misses        531      527       -4     
+ Partials     2656     2652       -4     

Flag	Coverage Δ
eest-develop	80.55% <100.00%> (+0.53%)	⬆️
eest-develop-gmp	16.12% <100.00%> (+0.19%)	⬆️
eest-legacy	11.18% <0.00%> (-0.01%)	⬇️
eest-legacy-silkpre	17.60% <0.00%> (-0.02%)	⬇️
evmone-unittests	83.75% <0.00%> (-0.06%)	⬇️

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

Components	Coverage Δ
core	93.74% <100.00%> (+0.12%)	⬆️
tooling	88.01% <ø> (ø)
tests	84.16% <ø> (ø)

Files with missing lines	Coverage Δ
test/state/precompiles.cpp	95.91% <100.00%> (+0.13%)	⬆️

... and 1 file with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[chfast]

I did something similar in #1046 for performance reasons, but there are additional details about BLST which complicates this. I think we should just remove procedures for single point multiplication if not used anywhere else.

[pdobacz]

I did something similar in #1046 for performance reasons, but there are additional details about BLST which complicates this. I think we should just remove procedures for single point multiplication if not used anywhere else.

Thx! For posterity, can we make note here of the complications? I couldn't find by quickly scanning the linked content, and the EIP is quite explicit about its recommendations.

OTOH: if the EIP says what it says, and we don't have a dedicated function in evmone, and yet we're measuring and reasoning about coverage using evmone, this approach becomes a bit inadequate - we could be missing coverage for the single multiplication case and not see it, while other clients might have been running (uncovered) single multiplication code. As a minimum, should we at least have a trivial input_size if, like in this PR, but both legs dispatching to the MSMs?

[chfast]

Fine to keep it.