RFR: 8365407: Race condition in MethodTrainingData::verify() [v4]

[Igor Veresov]

On Thu, 21 Aug 2025 16:59:54 GMT, Vladimir Kozlov <kvn at openjdk.org> wrote:

>> Igor Veresov has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   More cleanup
>
> src/hotspot/share/compiler/compilationPolicy.cpp line 143:
> 
>> 141:  void CompilationPolicy::wait_replay_training_at_init(JavaThread* THREAD) {
>> 142:     MonitorLocker locker(THREAD, TrainingReplayQueue_lock);
>> 143:     while (!_training_replay_queue.is_empty_unlocked() || _training_replay_queue.is_processing_unlocked()) {
> 
> Is this queue used in all phases?

For now just in production and assembly. During training the queue is there but is empty. But with iterative training it's going to be present during training.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/26866#discussion_r2291714279