On Mon, 14 Sep 2026 21:11:10 GMT, Vladimir Ivanov <[email protected]> wrote:

> Do we have any test cases to check deoptimization actions in presence of 
> concurrent related/unrelated closing actions? (E.g., if unrelated segment is 
> closed, there's no deoptimization action observed. And the opposite.)

We have 
https://github.com/openjdk/jdk/blob/master/test/micro/org/openjdk/bench/java/lang/foreign/ConcurrentClose.java
 to measure the impact of an unrelated shared arena being closed concurrently.

https://github.com/openjdk/jdk/blob/master/test/jdk/java/foreign/TestHandshake.java
 was used to shake out issues with missing deopts.

It is tricky to reliably catch a thread in the middle of an access, but I think 
it should be possible to do using JVMTI. I'll see if I can add a more explicit 
test that checks for deopts as well.

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

PR Comment: https://git.openjdk.org/jdk/pull/30926#issuecomment-5671126891

Reply via email to