RFR: 8241478: vmTestbase/gc/gctests/Steal/steal001/steal001.java fails with OOME
Kim Barrett
kim.barrett at oracle.com
Thu Mar 26 19:32:33 UTC 2020
> On Mar 26, 2020, at 3:25 PM, Thomas Schatzl <thomas.schatzl at oracle.com> wrote:
>
> Hi,
>
> On 26.03.20 18:46, Leonid Mesnik wrote:
>> Hi
>> Could you please review following fix which removes vmTestbase/gc/gctests/Steal tests from the repo.
>> These tests might fail provoking OutOfMemoryError throwing it in unexpected place. The logic and intention of these tests are very unclear from their description and code.
>> Test allocate objects till OOME. Than test removes some references ans start allocate objects provoking OOME and corresponding GC. It doesn't clear how it should stress taskqueue work-stealing in GC better than any other test causing a lot of GC.
>> This test however mentioned in a couple of bugs where crashes were caused by OOME in Heap.
>> https://bugs.openjdk.java.net/browse/JDK-8180627
>> https://bugs.openjdk.java.net/browse/JDK-8130344
>
> Please close these out as duplicates.
>
>> Assuming that we already have enough tests stressing OOME and tests stressing GC. I think it makes sense just to remove these two tests.
>
> Some analysis of work stealing statistics shows that other existing tests stress work stealing more than these two.
>
>> I verified that they are not in any jtreg group.
>> webrev: http://cr.openjdk.java.net/~lmesnik/8241478/webrev.00/
>> bug: https://bugs.openjdk.java.net/browse/JDK-8241478
>
> looks good. Thanks.
>
> Thomas
Looks good to me too.
More information about the hotspot-gc-dev
mailing list