fix: splitstore: remove deadlock around waiting for sync

[arajasek]

Related Issues

#10788?

#10641 introduced the possibility of a deadlock, because Put might sleep waiting on sync to catch up while holding the all-important txnLk. This is definitely wrong, we also just don't want Puts to be yielding and sleeping.

Proposed Changes

Add a non-yielding version of walkObjectIncomplete that gets called by Put and PutMany.

Additional Info

I checked thoroughly, but am still not confident that other methods more directly involved in compaction don't have the same problem. This entire locking mechanism is...tricky.

Checklist

Before you mark the PR ready for review, please make sure that:

• Commits have a clear commit message.
• PR title is in the form of of <PR type>: <area>: <change being made>
  • example: fix: mempool: Introduce a cache for valid signatures
  • PR type: fix, feat, build, chore, ci, docs, perf, refactor, revert, style, test
  • area, e.g. api, chain, state, market, mempool, multisig, networking, paych, proving, sealing, wallet, deps
• New features have usage guidelines and / or documentation updates in
  • Lotus Documentation
  • Discussion Tutorials
• Tests exist for new functionality or change in behavior
• CI is green

[ZenGround0]

Another approach that is a bit cleaner is to change this (and maybe walkObject) to always call checkClosing, then do a checkYield in every call to walkBlock. Its probably fine for sync protection to just do yielding once for every block walked during compaction.

[arajasek]

Superseded by #10857 because GH confused