serviceability_jvmti_j9_1_FAILED serviceability/jvmti/thread/GetStackTrace/getstacktr07/getstacktr07.java Fatal error: SetBreakpoint failed

[JasonFengJ9]

Failure link

From an internal build(macaarch64rt1):

openjdk version "22.0.1-beta" 2024-04-16
IBM Semeru Runtime Open Edition 22.0.1+8-202406120002 (build 22.0.1-beta+8-202406120002)
Eclipse OpenJ9 VM 22.0.1+8-202406120002 (build master-58b586d83e, JRE 22 Mac OS X aarch64-64-Bit 20240612_88 (JIT enabled, AOT enabled)
OpenJ9   - 58b586d83e
OMR      - 873509033
JCL      - ae28e24bf based on jdk-22.0.1+8)

Rerun in Grinder - Change TARGET to run only the failed test targets.

Optional info

Failure output (captured from console output)

[2024-06-12T09:24:22.047Z] variation: Mode650
[2024-06-12T09:24:22.047Z] JVM_OPTIONS:  -XX:-UseCompressedOops -Xverbosegclog 

[2024-06-12T09:25:16.669Z] TEST: serviceability/jvmti/thread/GetStackTrace/getstacktr07/getstacktr07.java

[2024-06-12T09:25:16.671Z] STDERR:
[2024-06-12T09:25:16.671Z] JVMVMUT015W Warning: Extended HCR is disabled, enable with -XX:+EnableExtendedHCR.
[2024-06-12T09:25:16.671Z] 
[2024-06-12T09:25:16.671Z] Fatal error: SetBreakpoint failed.

[2024-06-12T09:26:27.240Z] serviceability_jvmti_j9_1_FAILED

It is related to

• Add cmdline opts to toggle extendedHCR #19554

FYI @tajila

[tajila]

@JasonFengJ9 @LongyuZhang Do either of you know home to modify the options we run jtreg tests with?

[JasonFengJ9]

Assuming this test needs to run with -XX:+EnableExtendedHCR, @tajila I will add it.

[pshipton]

Options can be added to individual tests.
https://github.com/ibmruntimes/openj9-openjdk-jdk22/blob/openj9/test/hotspot/jtreg/serviceability/jvmti/thread/GetStackTrace/getstacktr07/getstacktr07.java#L42

Options can be added when we run the entire bucket.
https://github.com/adoptium/aqa-tests/blob/master/openjdk/playlist.xml#L409

Isn't it bad that something which is expected to work isn't working without the option?

[pshipton]

Test description

 *     The test exercises JVMTI function GetStackTrace via breakpoint
 *     and consequent class redefinition.
 *     The test starts a new thread, does some nested calls, stops at breakpoint,
 *     redefines the class and checks if the number of frames in the thread's
 *     stack is as expected and the function returns all the expected frames.

@requires vm.continuations

[tajila]

Assuming this test needs to run with -XX:+EnableExtendedHCR, @tajila I will add it.

Thanks

[JasonFengJ9]

Opened a PR

• serviceability_jvmti_j9 runs with -XX:+EnableExtendedHCR adoptium/aqa-tests#5385

[pshipton]

@tajila pls explain why it's ok to ignore the failing test.

Do we really want to run all the serviceability tests with -XX:+EnableExtendedHCR, which hides any problems caused by turning off the feature by default?

[tajila]

I didn't actually look too closely at the test. I assumed it was doing something that required extended capabilities. Ill take another look at it.

[tajila]

@JasonFengJ9 did this only fail in JDK22?

[JasonFengJ9]

did this only fail in JDK22?

@tajila serviceability_jvmti_j9 is enabled for JDK 21+, and it failed at JDK21 aarch64_linux as well