Init sync arbitrary batch length + full bandwidth utilization

[farazdagi]

What type of PR is this?

Refactoring+Optimization

What does this PR do? Why is it needed?

• Allows to utilize bandwidth efficiently by switching from epoch-based queue (where each epoch was assigned a FSM), to batches of arbitrary length (currently 64 - to match req/sec params).
• Removes redundant code (in FSM mostly).

Relevant PR

• Re-check issue Block is invalid error during initial sync #5976 after this is merged.

Other notes for review

• part of Init sync FSM, weighted RR, full-bandwidth optimizations #5798
• This is almost the final PR within the current optimization round (concluding PR will have more tests to increase coverage - no functionality will be added). So, with --dev flag, when syncing from genesis, you'll will occasionally see around 100blks/sec. Overall, speed of syncing is ~70-80blks/sec.
• Less requests needed to sync the same amount of chain blocks (previously we were restricted to 32 blocks/req, now it is arbitrary, and set to 64blk/req, so to sync the same chain we need half as many requests as previously)
• Since core underlying data structure is changed in both queue and FSMs, I've tried to simplify and improve existing code benefiting from the knowledge about syncing I possess now (re-visiting init-sync was quite a fun project).
• During last couple of days, I've synced dozens of times (with and without the --dev flag).

[codecov]

Codecov Report

Merging #5978 into master will decrease coverage by 0.34%.
The diff coverage is 82.38%.

@@            Coverage Diff             @@
##           master    #5978      +/-   ##
==========================================
- Coverage   59.89%   59.54%   -0.35%     
==========================================
  Files         321      320       -1     
  Lines       27632    26991     -641     
==========================================
- Hits        16549    16073     -476     
+ Misses       8866     8731     -135     
+ Partials     2217     2187      -30