Two (since 19) usages of Unsafe.getAndAddInt to update a static field provide
the Class object as the base instead of the base object returned by
Unsafe.staticFieldBase. This doesn't change anything on the HotSpot VM.
-
Commit messages:
- Initial commit
Changes: https://git.openjd
On Fri, 25 Nov 2022 14:53:57 GMT, Alan Bateman wrote:
> Two (since 19) usages of Unsafe.getAndAddInt to update a static field provide
> the Class object as the base instead of the base object returned by
> Unsafe.staticFieldBase. This doesn't change anything on the HotSpot VM.
LGTM
--
On Fri, 25 Nov 2022 14:53:57 GMT, Alan Bateman wrote:
> Two (since 19) usages of Unsafe.getAndAddInt to update a static field provide
> the Class object as the base instead of the base object returned by
> Unsafe.staticFieldBase. This doesn't change anything on the HotSpot VM.
LGTM, thank you!
On Tue, 29 Nov 2022 16:26:54 GMT, Bernhard Urban-Forster
wrote:
>> Two (since 19) usages of Unsafe.getAndAddInt to update a static field
>> provide the Class object as the base instead of the base object returned by
>> Unsafe.staticFieldBase. This doesn't change anything on the HotSpot VM.
>
>
On Fri, 25 Nov 2022 14:53:57 GMT, Alan Bateman wrote:
> Two (since 19) usages of Unsafe.getAndAddInt to update a static field provide
> the Class object as the base instead of the base object returned by
> Unsafe.staticFieldBase. This doesn't change anything on the HotSpot VM.
The introduction
On Wed, 30 Nov 2022 02:19:15 GMT, Martin Buchholz wrote:
> The introduction of POOLIDS_BASE strengthens the case for renaming POOLIDS to
> POOLIDS_OFFSET for clarity - I would do that.
I checked with Doug and he suggests we leave it as is.
-
PR: https://git.openjdk.org/jdk/pull/11