RFR: 8310340: assert(_thread->is_interp_only_mode() || stub_caller) failed: expected a stub-caller

[Patricio Chilano Mateo]

On Mon, 6 Jan 2025 21:20:28 GMT, David Holmes <dholmes at openjdk.org> wrote:

>> Please review the following fix. In method `JvmtiEventControllerPrivate::recompute_thread_enabled()`, we are missing to call `leave_interp_only_mode()` for the case where `should_be_interp` is computed as false and `state->is_pending_interp_only_mode()` is true. I added the full trace leading to the crash in the bug comments.
>> In JDK-8338383 I removed this assert because the branch condition changed and it became sort of a redundant check. But given that it was able to find this issue I added it back.
>> I was able to reproduce the crash easily by adding an extra delay before the assert. I verified the crash doesn’t reproduce anymore with this fix. I also run the patch through mach5 tiers 1-7.
>> 
>> Thanks,
>> Patricio
>
> Looks good to me. Thanks for the detailed analysis.

Thanks for the reviews @dholmes-ora, @alexmenkov and @sspitsyn!

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

PR Comment: https://git.openjdk.org/jdk/pull/22931#issuecomment-2578734644