RFR: 8339889: Several compiler tests ignore vm flags and not marked as flagless [v2]

[Damon Fenacci]

On Tue, 11 Feb 2025 22:43:35 GMT, Leonid Mesnik <lmesnik at openjdk.org> wrote:

>> Tests
>> compiler/c2/TestReduceAllocationAndHeapDump.java
>> compiler/calls/NativeCalls.java
>> compiler/debug/TestStress.java
>> compiler/inlining/TestDuplicatedLateInliningOutput.java
>> ignore vm flags using limited process builder and not marked as flagless.
>> 
>> Please note that test
>> compiler/inlining/TestDuplicatedLateInliningOutput.java
>> is failing with some VM flags. See
>> https://bugs.openjdk.org/browse/JDK-8348214
>> 
>> I haven't excluded test, since it fail with certain non-common flags only.
>
> Leonid Mesnik has updated the pull request incrementally with one additional commit since the last revision:
> 
>   test updated as suggested.

Thanks for "cleaning this up" @lmesnik.

I just ran a quick grep on `test/hotspot/jtreg/compiler` and noticed that there are a few more tests that use `ProcessTools.createLimitedTestJavaProcessBuilder` but don't have `vm.flagless` and don't seem to be covered by other JBS issues (e.g. `compiler/codecache/CheckLargePages.java`, `compiler/onSpinWait/TestOnSpinWaitAArch64DefaultFlags.java`, `compiler/jvmci/TestUncaughtErrorInCompileMethod.java` or `compiler/jvmci/compilerToVM/GetFlagValueTest.java`). Their main method runs in a new VM (`@run main/othervm`) but then they run other processes with `ProcessTools.createLimitedTestJavaProcessBuilder `. As I understand, vm flags would only affect the main method (which supposedly is not what is being tested). So, I was wondering if it made sense to mark them flagless as well anyway.

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

PR Comment: https://git.openjdk.org/jdk/pull/23224#issuecomment-2662378337