RFR: 8251358: Clean up Access configuration after Shenandoah barrier change [v2]

Aditya Mandaleeka adityam at openjdk.java.net
Mon Sep 28 20:35:26 UTC 2020


On Thu, 24 Sep 2020 07:27:17 GMT, Erik Österlund <eosterlund at openjdk.org> wrote:

>> Oops seems like I didn't publish the comments somehow.
>
>> Thanks @shipilev and @fisk . I had left `cas_set_mark_raw` alone since there's a caller that specifies
>> `memory_order_release` and I didn't want to change any behavior with the cleanup:
>> https://github.com/openjdk/jdk/blob/master/src/hotspot/share/gc/parallel/psPromotionManager.inline.hpp#L242
>> 
>> That said, I haven't looked into how necessary that is; perhaps it's okay to change it to relaxed order.
>> 
>> 
>> 
>> @fisk Thanks for taking another look. I don't see the new comments you added yet. Have you submitted them?
>> 
>> 
>> 
>> I'll follow up on this PR tomorrow since it's quite late here ��
>> 
>> 
> 
> A lot of thought has gone into the memory ordering flags. This change should definitely not change any of that. Having
> said that, I can't see why we can not rename raw functions to not raw functions, while maintaining the same memory
> ordering.

@fisk Can you please take another look at this when you get a chance?

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

PR: https://git.openjdk.java.net/jdk/pull/146


More information about the shenandoah-dev mailing list