RFR: 8300708: some nsk jvmti tests fail with virtual thread wrapper due to jvmti missing some virtual thread support

[Chris Plummer]

On Mon, 27 Jan 2025 11:24:33 GMT, Serguei Spitsyn <sspitsyn at openjdk.org> wrote:

> Some `nsk.jvmti` tests are failed with the `-testThreadFactory:Virtual`, so they are listed in the `test/hotspot/jtreg/ProblemList-Virtual.txt`.
> 
> The fix is to adjust tests expectations regarding tested virtual threads to expect error `JVMTI_ERROR_OPAQUE_FRAME`.
> 
> Testing:
> - Ran mach5 tiers 1-5

test/hotspot/jtreg/ProblemList-Virtual.txt line 41:

> 39: 
> 40: vmTestbase/nsk/jvmti/GetCurrentThreadCpuTime/curthrcputime001/TestDescription.java 8300708 generic-all
> 41: vmTestbase/nsk/jvmti/GetThreadCpuTime/thrcputime001/TestDescription.java 8300708 generic-all

You still have these two unfixed tests in the problem list using the CR for this PR. Either they will need a new CR, or you will need to use a different CR for this PR.

test/hotspot/jtreg/vmTestbase/nsk/jvmti/SetLocalVariable/setlocal001.java line 105:

> 103:             checkPoint();
> 104:             if (currThread.isVirtual()) {
> 105:                 out.println("meth01: skipping results check for virtual thread");

Should be "meth02". Similar issues below.

test/hotspot/jtreg/vmTestbase/nsk/jvmti/SetLocalVariable/setlocal001/setlocal001.cpp line 78:

> 76:                 err = jvmti_env->SetLocalLong(thr, 1,
> 77:                     table[i].slot, longVal);
> 78:                 if (err != JVMTI_ERROR_NONE && !(is_virtual && err == JVMTI_ERROR_OPAQUE_FRAME)) {

Maybe add a new API called something `invalidError(err)`. Not necessary, but might be a little better.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/23314#discussion_r1931203559
PR Review Comment: https://git.openjdk.org/jdk/pull/23314#discussion_r1931207856
PR Review Comment: https://git.openjdk.org/jdk/pull/23314#discussion_r1931210124