On Tue, 24 Feb 2026 19:28:51 GMT, Serguei Spitsyn <[email protected]> wrote:
>> The `interp-only` mechanism is based on the `JavaThread` objects. Carrier
>> and virtual threads can temporary share the same `JavaThread`. The
>> `java_thread->jvmti_thread_state()` is re-linked to a virtual thread at
>> `mount` and to the carrier thread at `unmount`. The `JvmtiThreadState` has a
>> back link to the `JavaThread` which is also set for virtual thread at a
>> `mount` and carrier thread at an `unmount`. Just one of these two links at
>> the same time is set to the `JavaThread`, the other one has to be set to
>> `nullptr`. The `interp-only` mechanism needs this invariant.
>> However, there is a corner case when this invariant is broken. It happens
>> when the `JvmtiThreadState` for carrier thread has just been created. In
>> such case, the link to `JavaThread` is always `non-nullptr` even though a
>> virtual thread is currently mounted on a carrier thread. This simple update
>> fixes the issue in the `JvmtiThreadState` ctor.
>>
>> Also, this update the includes the `interp_only` implementation
>> simplifications and more asserts are added to relevant places. One of the
>> simplification is a removal of the field `JvmtiThreadState::_thread_saved`.
>>
>> Testing:
>> - TBD: Mach5 tiers 1-6
>
> Serguei Spitsyn has updated the pull request incrementally with one
> additional commit since the last revision:
>
> review: convert a couple of sanity checks into asserts
src/hotspot/share/prims/jvmtiEventController.cpp line 381:
> 379: if (target == nullptr ||
> // an unmounted virtual thread
> 380: JvmtiEnvBase::is_thread_carrying_vthread(target,
> state->get_thread_oop())) { // a vthread carrying thread
> 381: return; // EnterInterpOnlyModeClosure will be executed right after
> mount.
The condition is not clear to me. How we do the deoptimization for mounted
vthread?
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/29436#discussion_r2855608966
