RFR: 8283276: java/io/ObjectStreamClass/ObjectStreamClassCaching.java fails with various GCs

Peter Levart plevart at openjdk.org
Fri Jul 29 10:39:57 UTC 2022


On Fri, 29 Jul 2022 09:05:53 GMT, Peter Levart <plevart at openjdk.org> wrote:

> This is a continuation of effort from https://github.com/openjdk/jdk/pull/9533 to fix the ObjectStreamClassCaching test which is failing with various GCs != G1. The test class contains 2 test methods:
> - test2CacheReleaseUnderMemoryPressure - this one was not logically changed at all - just one method was inlined
> - test1CacheEffectiveness - this one now uses a different strategy which doesn't involve calling System.gc() in order to trigger reference processing in GC which is, as test failures reveal, sometimes to aggressive and triggers processing not only WeakReference(s) but also SoftReference(s). Instead, the test now gradually builds up memory pressure while checking what's happening to two WeakReference(s): ref1 - wrapping a cached ObjectStreamClass instance; and: ref2 - wrapping a new Object() instance. The "effectiveness" of caching is confirmed by verifying that weakly reachable new Object() referent of ref2 is GC-ed earlier than softly reachable ObjectStreamClass referent of ref1.
> The test now contains several @run(s) with explicitly selected set of GC algorithms: G1, Parallel, ZGC, Shenandoah.

I can make another PR/issue to just problem-list the test if you feel that reviewing this PR might take some time. But as this is a test-only fix with noreg-self label and is intended to be backported to JDK 19, this would just duplicate the administrative effort...

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

PR: https://git.openjdk.org/jdk/pull/9684


More information about the core-libs-dev mailing list