some basic splitstore refactors

[vyzo]

Preliminaries for implementing additional functionality on top (sortless compaction, gc, etc), this performs some basic refactoring to avoid polluting ballooning subsequent prs:

• The distinction between MarkSet and MarkSetVisitor is removed.
• SetConcurrent is removed from MarkSet interface; all mark sets must be concurrent, because ...
• Chain walking is now parallelized

[codecov]

Codecov Report

Merging #7999 (f07ce29) into master (a246865) will decrease coverage by 0.09%.
The diff coverage is 74.71%.

@@            Coverage Diff             @@
##           master    #7999      +/-   ##
==========================================
- Coverage   39.19%   39.10%   -0.10%     
==========================================
  Files         660      660              
  Lines       71411    71428      +17     
==========================================
- Hits        27993    27932      -61     
- Misses      38598    38684      +86     
+ Partials     4820     4812       -8     

Impacted Files	Coverage Δ
blockstore/splitstore/markset.go	75.00% <ø> (ø)
blockstore/splitstore/splitstore.go	27.09% <ø> (+0.21%)	⬆️
blockstore/splitstore/splitstore_check.go	0.00% <0.00%> (ø)
blockstore/splitstore/splitstore_compact.go	52.43% <77.77%> (+0.67%)	⬆️
blockstore/splitstore/splitstore_warmup.go	52.63% <84.61%> (+1.24%)	⬆️
blockstore/splitstore/markset_badger.go	68.07% <100.00%> (-1.52%)	⬇️
blockstore/splitstore/markset_map.go	83.78% <100.00%> (+2.53%)	⬆️
blockstore/splitstore/visitor.go	100.00% <100.00%> (ø)
itests/kit/blockminer.go	74.11% <0.00%> (-17.06%)	⬇️
markets/loggers/loggers.go	88.88% <0.00%> (-11.12%)	⬇️
... and 26 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update a246865...f07ce29. Read the comment docs.

[magik6k]

Don't see anything obviously wrong, just one tiny nitpick

[magik6k]

Technically we don't need this copy; walking could be set to nil after populating workch

[vyzo]

ah yes, and we can reuse the slice; let me fix.

[vyzo]

hrm, we can't actually reuse the slice, it first comes from the tipset; will just setr to nil after copying.

[vyzo]

done.

[vyzo]

you know, we can first make a new one and copy when we initialize and then reuse; that will save some allocations.
let me do.

[vyzo]

done done.