RFR: 8329790: GenShen: Old regions that are pinned during final mark are not made parseable [v2]

[William Kemper]

On Sat, 6 Apr 2024 00:43:05 GMT, Kelvin Nilsen <kdnilsen at openjdk.org> wrote:

>> William Kemper has updated the pull request with a new target base due to a merge or a rebase. The incremental webrev excludes the unrelated changes brought in by the merge/rebase. The pull request contains six additional commits since the last revision:
>> 
>>  - Merge remote-tracking branch 'shenandoah/master' into coalesce-old-pinned-regions
>>  - Use is_regular_pinned to restore candidate selection logic to its previous form
>>  - Introduce 'regular pinned' state
>>  - Merge remote-tracking branch 'shenandoah/master' into coalesce-old-pinned-regions
>>  - Allow pinned/regular regions into mixed collection candidates
>>  - Little clean up
>
> src/hotspot/share/gc/shenandoah/heuristics/shenandoahOldHeuristics.cpp line 469:
> 
>> 467:   log_info(gc)("Old-Gen Immediate Garbage: " PROPERFMT " over " SIZE_FORMAT " regions",
>> 468:               PROPERFMTARGS(immediate_garbage), immediate_regions);
>> 469:   log_info(gc)("Old regions selected for defragmentation: " SIZE_FORMAT, defrag_count);
> 
> Might be good to add a comment here that defrag_count is a subset of old_candidates.  defrag_count represents regions that were placed into the old collection set in order to defragment the memory that we try to "reserve" for humongous allocations.  old_candidates includes defrag_count, plus it also includes old-gen regions that are to be collected because they contain an abundance of garbage.

Added a comment.

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

PR Review Comment: https://git.openjdk.org/shenandoah/pull/420#discussion_r1563314072