RFR: 8367717: Cleanup atomic_copy64
jonghoonpark
duke at openjdk.org
Mon Oct 6 07:50:57 UTC 2025
On Mon, 6 Oct 2025 06:44:48 GMT, Kim Barrett <kbarrett at openjdk.org> wrote:
>> related jira issue: https://bugs.openjdk.org/browse/JDK-8367717
>>
>> ## Changes
>> As described in the issue, this change aligns `atomic_copy64` definitions for zero platforms with those of other platforms.
>>
>> ## Verification
>> - Verified by running: `make run-test TEST="gtest:AtomicAccess*"` with a JDK image built using the `zero` JVM variant.
>>
>> Please let me know if there are any additional or more appropriate verification steps I should perform.
>
>> As described in the issue, this change aligns atomic_copy64 definitions for
>> zero platforms with those of other platforms.
>
> The issue suggests it might be better to replace calls to atomic_copy64 with
> `AtomicAccess::store(dst, AtomicAccess::load(src))` (using the new
> `AtomicAccess` nomenclature), and then delete all definitions of
> atomic_copy64. And there's a comment agreeing with that suggestion. I think
> that's the way to go, rather than what this PR proposes.
@kimbarrett
Thanks for the feedback.
I thought the plan was to align the definitions first, and then implement the new form later.
I'll revise the work to make the change to the new formulation, as you suggested.
-------------
PR Comment: https://git.openjdk.org/jdk/pull/27642#issuecomment-3370333776
More information about the hotspot-runtime-dev
mailing list