On Wed, 6 Sep 2023 20:02:44 GMT, Alex Menkov <amen...@openjdk.org> wrote:
> To test ForceEarlyReturn command for NO_MORE_FRAMES case the test creates > ThreadStartEventRequest with SUSPEND_ALL policy and requests debuggee to > start new thread. > If debuggee JVM starts some internal threads before the request is cleared > (i.e. we have several ThreadStart events), 2nd event suspends debuggee again > and the test fails with timeout. > The change adds THREAD_ONLY modifier to the ThreadStartEventRequest to > generate event only for desired thread. > It requires thread ID, so debuggee was updated to create Thread object in > advance, debugger reads the thread ID from static field (it does not need to > be static, but Debugee class has convenient methods to retrieve class ID and > static field value). > > Testing: 100 runs of the test on > windows-x64-debug,linux-x64-debug,macosx-x64-debug with > JTREG_TEST_THREAD_FACTORY=Virtual, with and without "-XX:+UseZGC > -XX:+ZGenerational" Overall looks good. Just one minor comment suggestion. test/hotspot/jtreg/vmTestbase/nsk/jdwp/ThreadReference/ForceEarlyReturn/forceEarlyReturn002/forceEarlyReturn002.java line 164: > 162: command.addByte(JDWP.EventKind.THREAD_START); > 163: command.addByte(JDWP.SuspendPolicy.ALL); > 164: // THREAD_ONLY modifier Maybe add a bit more of a comment here that explains we only want the THREAD_START event for the specified test thread and not any others that might start up. ------------- Marked as reviewed by cjplummer (Reviewer). PR Review: https://git.openjdk.org/jdk/pull/15601#pullrequestreview-1614398877 PR Review Comment: https://git.openjdk.org/jdk/pull/15601#discussion_r1317959286
