Request for Review (s) - 8060463: ParNew: SurvivorAlignmentInBytes greater then YoungPLABSize cause assert(obj != NULL | (original) (raw)

Request for Review (s) - 8060463: ParNew: SurvivorAlignmentInBytes greater then YoungPLABSize cause assert(obj != NULL || plab->words_remaining() < word_sz) failed: Else should have been able to allocate

Jon Masamitsu jon.masamitsu at oracle.com
Thu Mar 24 15:44:16 UTC 2016

Michail,

Thanks for reviewing this.

On 3/24/2016 5:33 AM, Michail Chernov wrote:

Hi Jon,

Two minor comments: 1. Test should have @requires because you use specified GC and ExplicitGCInvokesConcurrent. Our nightly testing uses ExplicitGCInvokesConcurrent, so there can be options conflict. Requires can be next: @requires (vm.gc == "ConcMarkSweep" | vm.gc == null) & vm.opt.ExplicitGCInvokesConcurrent != true

It occurs to me now that this test should run with all the GC's. Though the fix is only for CMS, using the large SurvivorAlignmentInBytes should work with the other GC's. I'll do a little more testing and then publish a new webrev without the -XX:+UseConcMarkSweepGC. I'll add the

@requires vm.opt.ExplicitGCInvokesConcurrent != true

2. Test does not need any @modules or @library because it does not use such modules/libraries.

Fixed.

Jon

Thanks Michail

On 03/21/2016 06:04 AM, Jon Masamitsu wrote:

https://bugs.openjdk.java.net/browse/JDK-8060463 http://cr.openjdk.java.net/~jmasa/8060463/webrev.00/

More information about the hotspot-gc-dev mailing list