[jdk17u-dev] RFR: 8227060: Optimize safepoint cleanup subtask order [v2]

[Aleksey Shipilev]

On Tue, 23 May 2023 09:02:18 GMT, Aleksey Shipilev <shade at openjdk.org> wrote:

>> Unclean backport to improve the safepoint cleanup times. Provides the grounds for the ultimate improvement in safepoint cleanup ([JDK-8306738](https://bugs.openjdk.org/browse/JDK-8306738)). The backport is not clean, because there are copyright years conflicts, the rename of `AbstractGangTask` to `WorkerTask` is missing, `WorkGang` -> `WorkerThreads` rename is missing, `JavaThread::cast` cleanup is missing.
>> 
>> Additional testing:
>>  - [x] Ad-hoc benchmarks
>>  - [x] Linux x86_64 fastdebug `tier1 tier2 tier3`
>> 
>> The safepoint cleanup times are slightly better on small tests, and there seem to be no regressions.
>> ![plot](https://github.com/openjdk/jdk17u-dev/assets/1858943/7e012482-71f8-459f-aca0-4c043b791e92)
>
> Aleksey Shipilev 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 three additional commits since the last revision:
> 
>  - Upstream hunk uses java_threads_do without is_JavaThread check
>  - Merge branch 'master' into JDK-8227060-optimize-safepoint-cleanup-order
>  - Backport 96a542feb2064dba155ebf05311752995d164038

> > Lgtm.
> 
> While looking at it again, I noticed that I missed the optimization from the original change: [1dbff5e](https://github.com/openjdk/jdk17u-dev/commit/1dbff5ed842e56b4620cd63f65be1a7f419930c7) -- please re-review. Re-doing testing again, expecting no problems.

@phohensee, if you can review again, please? Thanks!

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

PR Comment: https://git.openjdk.org/jdk17u-dev/pull/1377#issuecomment-1560719348