[jdk17u-dev] RFR: 8295724: VirtualMachineError: Out of space in CodeCache for method handle intrinsic [v2]

[Martin Doerr]

On Tue, 20 Dec 2022 22:42:52 GMT, Martin Doerr <mdoerr at openjdk.org> wrote:

>> Mostly clean backport of [JDK-8295724](https://bugs.openjdk.org/browse/JDK-8295724). Manual integration was needed because of unrelated changes in the surrounding code. There's only one significant difference:
>> The hunk from gcBehaviours.cpp needs to get integrated into `nmethod::is_unloading()` because the usage of `IsUnloadingBehaviour::current()` is there before [JDK-8290025](https://bugs.openjdk.org/browse/JDK-8290025).
>> Includes trivial fix for [JDK-8298947](https://bugs.openjdk.org/browse/JDK-8298947)
>
> Martin Doerr has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Backport 3d4d9fd6e6de037950f94482d4e33f178eb15daa

It was released with 17.0.7. Note that it only helps when -XX:+SegmentedCodeCache is enabled. And it still doesn't make the problem impossible when an insane amount of method handle intrinsics gets allocated. There are other solutions under discussion, but not yet implemented (https://bugs.openjdk.org/browse/JDK-8296336).
Also note that the backport contains a bug which is fixed for 17.0.8: https://bugs.openjdk.org/browse/JDK-8305088

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

PR Comment: https://git.openjdk.org/jdk17u-dev/pull/951#issuecomment-1546227748