RFR: 8343704: Bad GC parallelism with processing Cleaner queues [v5]

[Aleksey Shipilev]

On Thu, 14 Nov 2024 00:50:51 GMT, Chen Liang <liach at openjdk.org> wrote:

>> Aleksey Shipilev has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Review feedback: make sure trimming actually works, stylistic changes
>
> src/java.base/share/classes/jdk/internal/ref/CleanerImpl.java line 236:
> 
>> 234:     static final class PhantomCleanableList {
>> 235:         private static final int MIN_CAPACITY = 16;
>> 236:         private final Object lock = new Object();
> 
> Why use another lock object when identity of this list is not escaped elsewhere?

Mostly hygiene for not exposing the lock to external classes. But yeah, we can simplify by just doing `synchronized`. I'll push the update soon.

> src/java.base/share/classes/jdk/internal/ref/CleanerImpl.java line 311:
> 
>> 309:                 // want to cause an immediate resize on next insertion.
>> 310:                 if ((size < arr.length / 4) && (size > MIN_CAPACITY)) {
>> 311:                     int newLen = ArraysSupport.newLength(size, 1, size);
> 
> This code isn't immediately clear that it's trying to half the array size, as you are using the size computation method for array growth.  Can you add a remark for so?

None of this is actually needed once we do smarter thing than relying on array resizes (and copies). I'll push new implementation soon.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/22043#discussion_r1841897671
PR Review Comment: https://git.openjdk.org/jdk/pull/22043#discussion_r1841899531