Aleksey Shipilëv

Otherwise, we can review and try to integrate this with bots. Please take a look and approve this PR :)

Anyone? @iklam? :)

> Looks good to me in general. Can you comment why the changes are needed in deoptimization.cpp? Those are mostly moving the real counter code under `COMPILER2_OR_JVMCI` macro, leaving the...

> Thanks for the explanation. I've pushed your changes into premain: [b96c511](https://github.com/openjdk/leyden/commit/b96c511323aea6ffb2fc972817c779b4bacaf9fa) Thanks! And dang, I wanted to try and integrate this with bots! Oh well, next time.

Now waiting for https://github.com/openjdk/jdk/pull/21152 to arrive, so that we don't have a merge problem later.

Thanks for the merge, @iklam! I think we are ready to integrate this. /integrate

Please pull/rebase to current master :)

[JDK-8322846](https://bugs.java.com/bugdatabase/view_bug?bug_id=8322846) is already backported to 21.0.4 upstream, and would land in Corretto 21.0.4 release next week. That said, `java.lang.ClassCircularityError` does not look like [JDK-8322846](https://bugs.java.com/bugdatabase/view_bug?bug_id=8322846). `CCE` implicates the class transformation code...

GHA testing is still not enabled, please enable it and pass it to green.

/sponsor