RFR: Simplify G1CodeBlobClosure

Thomas Schatzl tschatzl at openjdk.java.net
Tue Mar 22 08:41:50 UTC 2022


On Tue, 22 Mar 2022 08:16:29 GMT, Stefan Karlsson <stefank at openjdk.org> wrote:

> There's no need to introduce a new _keepalive_is_strong property in the G1CodeBlobClosure. The existing _strong is set to true when the evacuation participates in the marking of the old gen, which is when we need to mark the nmethods as "maybe on continuation". The name _strong in this context is not great and it doesn't convey the actual meaning. Maybe something for a future G1 cleanup.

Thanks for looking into this - that change has been one of the stranger to me as well.

Looks good.

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

Marked as reviewed by tschatzl (no project role).

PR: https://git.openjdk.java.net/loom/pull/112


More information about the loom-dev mailing list