RFR: 8352251: Implement JEP 518: JFR Cooperative Sampling [v25]

[Markus Grönlund]

On Mon, 19 May 2025 14:06:21 GMT, Martin Doerr <mdoerr at openjdk.org> wrote:

> Thank you so much! This is probably very helpful. I had planned to do some experiments on my own, but I was too busy with other issues. I wonder if we really need memory barriers. The Java threads get suspended before we take a sample, right? So, I think the suspension mechanism should already prevent reordering of memory accesses. I'll try it on PPC64 as soon as I find some time.

I was going to ping you about the updates, but you saw them already :)

The initial store load barrier is to prevent the safepoint poll from floating up, the other is to ensure the fp is moved before the crit sec release. Defensive, I know.

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

PR Comment: https://git.openjdk.org/jdk/pull/24296#issuecomment-2891202291