RFR: JDK-8269393: store/load order not preserved when handling memory pool due to weakly ordered memory architecture of aarch64

[Andrew Haley]

On Fri, 6 Oct 2023 08:28:05 GMT, Andrew Haley <aph at openjdk.org> wrote:

>> @theRealAph do you think the current solution (with `CodeCache_lock` acquisition) could be OK?
>
> I think so. I am a bit nervous because `MutexLock` isn't recursive, so if this code is ever called from a region that is already locked it'll fail. But there is an assertion in `MutexLock` that will detect that if it ever happens, so OK.

Looks like I was right bout the locking, I'm afraid. We need a better way to handle this.

Still, at least the assertion detected this before everything went south.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/15819#discussion_r1348636409