Test8009761 fails on JDK 8 Windows 32-bit with Test8009761.m3(boolean,boolean) not compiled

[jiekang]

Seen during 2024 October CPU (#5692)

Test8009761 is failing on JDK 8 Windows 32-bit with logs:

[2024-10-17T12:39:14.876Z] STDOUT:
[2024-10-17T12:39:14.876Z] CompilerOracle: exclude Test8009761.m2
[2024-10-17T12:39:14.876Z] WB error: invalid compilation level 4
[2024-10-17T12:39:14.876Z] STDERR:
[2024-10-17T12:39:14.876Z] java.lang.RuntimeException: static java.lang.Object Test8009761.m3(boolean,boolean) not compiled
[2024-10-17T12:39:14.876Z] 	at Test8009761.main(Test8009761.java:263)
[2024-10-17T12:39:14.876Z] 	at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
[2024-10-17T12:39:14.876Z] 	at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
[2024-10-17T12:39:14.876Z] 	at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
[2024-10-17T12:39:14.876Z] 	at java.lang.reflect.Method.invoke(Method.java:498)
[2024-10-17T12:39:14.876Z] 	at com.sun.javatest.regtest.agent.MainWrapper$MainThread.run(MainWrapper.java:127)
[2024-10-17T12:39:14.876Z] 	at java.lang.Thread.run(Thread.java:750)

[jiekang]

This is a new test failure introduced with the backport of JDK-8021775

The new test code [1] is not Windows 32-bit friendly where C2 is not necessarily available.

[1]
https://github.com/openjdk/jdk8u/blob/618917eb093243de2c5d7e83d4688bfe9ad04985/hotspot/test/compiler/8009761/Test8009761.java#L260

Other test code [2] has constructs like:

            if(!WHITE_BOX.enqueueMethodForCompilation(m, COMP_LEVEL_FULL_OPTIMIZATION)) {
                // C2 compiler not available, try to compile with C1
                WHITE_BOX.enqueueMethodForCompilation(m, COMP_LEVEL_SIMPLE);
            }

To attempt compiler with C2 and then try C1 otherwise.

[2]
https://github.com/openjdk/jdk8u/blob/master/hotspot/test/compiler/classUnloading/anonymousClass/TestAnonymousClassUnloading.java#L88C1-L91C14

Related issue:
#5221

See also another test that failed in the same path:
#3045

[jiekang]

Based on above analysis, I would not consider this test failure a release blocker.