On Tue, 30 May 2023 19:32:11 GMT, Chris Plummer <cjplum...@openjdk.org> wrote:
>> Improve ThreadReference.ForceEarlyReturn to support virtual threads as long >> as they are suspended and mounted. >> >> [JDK-8308400](https://bugs.openjdk.org/browse/JDK-8308400) improved JVMTI >> ForceEarlyReturn support for virtual threads. The spec now says it is >> supported if the thread is suspended at an event (not a specific event as >> previously spec'd), and additional support may also be provided. The >> implementation now just requires that the virtual thread be mounted, so this >> covers being suspended at an event, but also other situations in which the >> virtual thread is mounted. >> >> JDWP and JDI need to line up with JVMTI. Only spec changes will be needed. >> >> A new test was added mostly to cover the unmounted virtual thread case that >> results in OpaqueFrameException. >> >> forceEarlyReturn002 test had previously been updated for virtual threads to >> expect OpaqueFrameException. These changes have all been undone since the >> test now runs the same when using virtual threads as with platform threads. > > Chris Plummer has updated the pull request incrementally with one additional > commit since the last revision: > > Restore forceEarlyReturn() platform thread behaviour to only throw > NativeMethodException and get rid of assert. A future CR/CSR will resolve the > issue of OPAQUE_FRAME sometimes not meaning NativeMethodException for > platform threads. The fix looks good. Thanks, Serguei ------------- Marked as reviewed by sspitsyn (Reviewer). PR Review: https://git.openjdk.org/jdk/pull/14167#pullrequestreview-1451722916
