RFR: 8305994: Guarantee eventual async monitor deflation [v2]

[Aleksey Shipilev]

On Fri, 14 Apr 2023 15:35:03 GMT, Volker Simonis <simonis at openjdk.org> wrote:

>> We don't know if this guaranteed deflation would make progress or not, so we cannot decrement blindly. In the new commit, I just introduced a new flag that disallows progress updates when we are in guaranteed deflation path.
>
> The new code fixes it , but isn't it unnecessarily complex? In the previous version, when we made progress, `_no_progress_cnt` would have been reset to zero anyway, and if we didn't made progress, we would have just incremented `_no_progress_cnt` which would have reverted the previous decrement. Or am I missing something?
> 
> Anyway, I'm fine with both versions.

The problem is here: https://github.com/openjdk/jdk/blob/1fd400608e9ea423a6e4f1797652bd02f027da1b/src/hotspot/share/runtime/synchronizer.cpp#L1072-L1073

Along with the reset of `_no_progress_cnt` in `monitors_used_above_threshold`, we *also* bump the ceiling, which means the next checks for used monitor thresholds get more and more unlikely to fire, because ceiling goes farther and farther away. You can clearly see that in the `-Xlog:monitordeflation` without the remediation like this: with every `NoAsyncDeflationProgressMax` guaranteed attempts, that block bumps the ceiling. :)

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/13474#discussion_r1166996886