RFR: 8252294: Remove OopsInGenClosure usage from younger_refs_iterate

Kim Barrett kim.barrett at oracle.com
Thu Aug 27 08:24:03 UTC 2020


> On Aug 27, 2020, at 3:48 AM, Stefan Karlsson <stefan.karlsson at oracle.com> wrote:
> 
> On 2020-08-27 04:36, Kim Barrett wrote:
>>> On Aug 25, 2020, at 4:12 AM, Stefan Karlsson <stefan.karlsson at oracle.com> wrote:
>>> 
>>> Hi all,
>>> 
>>> Please review this patch to remove the usage of OopsInGenClosure from younger_refs_iterate and related functions.
>>> 
>>> https://cr.openjdk.java.net/~stefank/8252294/webrev.01/
>>> https://bugs.openjdk.java.net/browse/JDK-8252294
>> Looks good.  Thanks for the detailed overview.
>> I don't need a new webrev for the comment deletion below.
>> ------------------------------------------------------------------------------
>> src/hotspot/share/gc/serial/serialHeap.cpp
>>  105   // When collection is parallel, all threads get to cooperate to do
>>  106   // old generation scanning.
>> This comment no longer makes any sense.
> 
> Removed.
> 
>> ------------------------------------------------------------------------------
>> src/hotspot/share/gc/serial/serialHeap.cpp
>>  101   if (_process_strong_tasks->try_claim_task(GCH_PS_younger_gens)) {
>> ...
>>  112   _process_strong_tasks->all_tasks_completed(scope->n_threads());
>> The use of parallel task support no longer makes sense here. I'm
>> guessing this is a cleanup you want to save for later?
> 
> Yeah. I didn't want to disturb the claiming code. But I'll remove it in this patch and rerun the tests.

I’m okay with deferring to a later cleanup.  Sorry if I wasn’t clear about that.

> 
> Thanks for reviewing,
> StefanK
> 
>> ------------------------------------------------------------------------------




More information about the hotspot-gc-dev mailing list