May 2013 - hotspot-gc-dev

RFR: 8013590: NPG: Add a memory pool MXBean for Metaspace
by Erik Helin 04 Jun '13

04 Jun '13

Hi all, this change adds two memory pools for metaspace, one for Metaspace and one for compressed klass space. The memory pool for compressed klass space will only have valus that differ from 0 or -1 (undefined) if compressed klass pointers are used. This change also adds a manager for the pools: Metaspace Manager. I have also added a test that checks that the metaspace manager is present and that the two pools are present. The test also verifies that the compressed klass space pool act correct according to the UseCompressedKlass flag. The last time I added metaspace memory pools, it triggered some unforeseen bugs: - Two asserts in jmm_GetMemoryUsage that asserted that a memory pool was either of heap type or had an undefined init/max size. - The jdk/test/java/lang/management/MemoryMXBean/MemoryTest.java failed - The service lock was taken out of order with the metaspace locks These bugs have all been fixed: - The asserts have been removed since they are no longer true - The test has been updated but is not part of this change since it is a JDK change - This change does not make use of functions requiring the metaspace lock. I had to remove some verification code in free_chunks_total to ensure this. Webrev: http://cr.openjdk.java.net/~ehelin/8013590/webrev.00/ Testing: - One new jtreg test - JPRT - All the tests that failed in nighly testing last time now pass Thanks, Erik

2 1

Re: Request for review: 6976350 G1: deal with fragmentation while copying objects during GC
by Tao Mao 04 Jun '13

04 Jun '13

To the open list: new webrev: http://cr.openjdk.java.net/~tamao/6976350/webrev.03/ I took suggestion from many reviewers into consideration and came up with the current cleaner solution. Thank you. Tao On 5/14/13 2:26 PM, Jon Masamitsu wrote: > What's the status of this review? > > The last mail I could find in my mail boxes was a comment > from Thomas. > > Jon > > On 1/28/13 12:21 PM, Tao Mao wrote: >> 6976350 G1: deal with fragmentation while copying objects during GC >> https://jbs.oracle.com/bugs/browse/JDK-6976350 >> >> webrev: >> http://cr.openjdk.java.net/~tamao/6976350/webrev.00/ >> >> changeset: >> Basically, we want to reuse more of par-allocation buffers instead of >> retiring it immediately when it encounters an object larger than its >> remaining part. >> >> (1) instead of previously using one allocation buffer per GC purpose, >> we use N(=2) buffers per GC purpose and modify the corresponding >> code. The changeset would easily scale up to whatever N (though Tony >> Printezis suggests 2, or 3 may be good enough) >> >> *(2) Two places of cleanup: allocate_during_gc_slow() is removed due >> to its never being called. >> access modifier >> (public) before trim_queue() is redundant. >> >> >

4 18

RFR (XSS): 8015735: TestObectCountAfterGCEvent.java should use targetObjectCountEvents instead of objectCountEvents
by Erik Helin 31 May '13

31 May '13

Hi all, could someone please review this tiny change: http://wikifiles.se.oracle.com/dev/ehelin/8015735/webrev.00/ The test jdk/test/closed/com/oracle/jfr/gc/TestObjectCountAfterGCEvent.java used the wrong list of object count events. It should use targetObjectCountEvents since these belong to just one particular GC (the list objectCountEvents might contain events from mulitple GC invocations). Thanks, Erik

1 0

hg: hsx/hotspot-gc/hotspot: 8015486: PSScavenge::is_obj_in_young is unnecessarily slow with UseCompressedOops
by stefan.karlsson＠oracle.com 31 May '13

31 May '13

Changeset: 47bdfb3d010f Author: stefank Date: 2013-05-30 10:58 +0200 URL: http://hg.openjdk.java.net/hsx/hotspot-gc/hotspot/rev/47bdfb3d010f 8015486: PSScavenge::is_obj_in_young is unnecessarily slow with UseCompressedOops Summary: Compare compressed oops to a compressed young gen boundary instead of uncompressing the oops before doing the young gen boundary check. Reviewed-by: brutisso, jmasa ! src/share/vm/gc_implementation/parallelScavenge/psPromotionManager.inline.hpp ! src/share/vm/gc_implementation/parallelScavenge/psScavenge.cpp ! src/share/vm/gc_implementation/parallelScavenge/psScavenge.hpp ! src/share/vm/gc_implementation/parallelScavenge/psScavenge.inline.hpp

1 0

Re: Request for review: 8009778: NPG: ClassMetaspaceSize is used before set in set_ergonomics_flags()
by Coleen Phillimore 31 May '13

31 May '13

Hi Thomas, It doesn't look like this change fixes the set before used problem of ClassMetaspaceSize but is a more comprehensive fix so I think you should fix Harold's bug with your fix. I only looked at arguments.cpp,hpp though. Actually, both changes are wrong for this because you only set ClassMetaspaceSize to 100*M (agreed, not a good size) if UseCompressedKlassPointers is set which you don't set until later. Harold is working on a fix move the class metaspace somewhere else, not attached to the Java heap calculations which will resolve this issue. So I think we should leave this broken for now. Coleen On 5/30/2013 4:59 PM, Thomas Schatzl wrote: > Hi all, > > On Thu, 2013-05-30 at 14:05 -0400, harold seigel wrote: >> Hi, >> >> Please review this fix for bug 8009778. >> >> The fix adds a function, class_metaspace_size_for_compressed_ptrs(), >> that returns what the value will be for ClassMetaspaceSize if >> CompressedKlassPointers are used. This enables the correct value for >> ClassMetaspaceSize to be used by max_heap_for_compressed_oops() before >> ClassMetaspaceSize gets set in set_ergonomics_flags(). >> > You may want to have a look at the change for 8010722 which is currently > under review on the hotspot-gc list, which (probably) also fixes this > one. Unfortunately I forgot to CC hotspot-runtime-dev. > > The mail at > http://mail.openjdk.java.net/pipermail/hotspot-gc-dev/2013-May/007155.html is the start of that thread; it would be nice btw if somebody of you did have some time to look over it. > > That change is very similar, I have no particular preference for either > solution. > > I have some comment about the choice of 100M for the default table size: > it is a very "odd" size for heap areas. I mean, later in > Universe::reserve_heap() this size is aligned up by the alignment given > to that function. > > That alignment is always a power of two, and there are a few cases when > this required alignment is quite big, i.e. 32M+. > > One case is if G1 selects a heap region size of 32M, the other much more > common is probably on SPARC where if the "small" OS page size is set to > 64k, this alignment is also always 32M regardless of collector. > > Hotspot further down actually expands the actual ClassMetaspaceSize > size. So it is no real problem except that it seemed somewhat strange to > me. This is just an observation, nothing to do about it (I think), I > just asked myself why somebody would choose a value of 100M when doing > the fix for 8010722 knowing that there's a lot of power-of-two > alignments in that code. > > Btw, I do think there's some issue here with > Universe::set_class_metaspace_size() and the alignment in > Universe::reserve_heap() though: the Universe::_class_metaspace_size > does not seem to be synchronized with the ClassMetaspaceSize global, so > any tool (jmap I think) that later reads and prints this value from the > global will give wrong output. I.e. return the unaligned, not the > actual, value at startup. > >> The fix was tested with JCK lang and vm tests, UTE vm.quick.testlist and >> vm.mlvm.testlist tests, JPRT, and jtreg tests. Tests were run on Linux >> and Solaris. The debugger was used to step through the code to ensure >> that the fix behaved as expected. >> >> Open webrev at http://cr.openjdk.java.net/~hseigel/bug_8009778/ >> <http://cr.openjdk.java.net/%7Ehseigel/bug_8009778/> >> >> Bug link at http://bugs.sun.com/view_bug.do?bug_id=8009778 >> >> JBS bug link at https://jbs.oracle.com/bugs/browse/JDK-8009778 > Apologies again for not CC'ing hotspot-runtime-dev about the 8010722. > issue. > > Thanks, > Thomas > >

1 0

hg: hsx/hotspot-gc/nashorn: Added tag jdk8-b92 for changeset dee23cce5235
by john.coomes＠oracle.com 31 May '13

31 May '13

Changeset: 1c7481ac7fe0 Author: katleman Date: 2013-05-30 10:58 -0700 URL: http://hg.openjdk.java.net/hsx/hotspot-gc/nashorn/rev/1c7481ac7fe0 Added tag jdk8-b92 for changeset dee23cce5235 ! .hgtags

1 0

hg: hsx/hotspot-gc/langtools: 4 new changesets
by john.coomes＠oracle.com 31 May '13

31 May '13

Changeset: 58eace4d997f Author: erikj Date: 2013-05-28 08:49 +0200 URL: http://hg.openjdk.java.net/hsx/hotspot-gc/langtools/rev/58eace4d997f 8012566: Replace find, rm, printf and similar with their proper variables Reviewed-by: tbell ! makefiles/BuildLangtools.gmk Changeset: 3597773628a4 Author: katleman Date: 2013-05-28 17:58 -0700 URL: http://hg.openjdk.java.net/hsx/hotspot-gc/langtools/rev/3597773628a4 Merge Changeset: 149890642a0e Author: katleman Date: 2013-05-29 10:16 -0700 URL: http://hg.openjdk.java.net/hsx/hotspot-gc/langtools/rev/149890642a0e 8015525: JDK8 b91 source with GPL header errors Reviewed-by: dholmes, lancea ! test/tools/javac/annotations/typeAnnotations/classfile/TestNewCastArray.java Changeset: 023e9a614d26 Author: katleman Date: 2013-05-30 10:58 -0700 URL: http://hg.openjdk.java.net/hsx/hotspot-gc/langtools/rev/023e9a614d26 Added tag jdk8-b92 for changeset 149890642a0e ! .hgtags

1 0

hg: hsx/hotspot-gc/jdk: 2 new changesets
by john.coomes＠oracle.com 31 May '13

31 May '13

Changeset: a2a2a91075ad Author: katleman Date: 2013-05-29 10:16 -0700 URL: http://hg.openjdk.java.net/hsx/hotspot-gc/jdk/rev/a2a2a91075ad 8015525: JDK8 b91 source with GPL header errors Reviewed-by: dholmes, lancea ! test/java/util/stream/test/org/openjdk/tests/java/util/FillableStringTest.java ! test/java/util/stream/test/org/openjdk/tests/java/util/MapTest.java ! test/jdk/lambda/MethodReferenceTestInstanceMethod.java ! test/jdk/lambda/MethodReferenceTestKinds.java ! test/jdk/lambda/MethodReferenceTestSueCase1.java ! test/jdk/lambda/MethodReferenceTestSueCase2.java ! test/jdk/lambda/MethodReferenceTestSueCase4.java ! test/jdk/lambda/separate/AttributeInjector.java ! test/jdk/lambda/separate/ClassFile.java ! test/jdk/lambda/separate/ClassFilePreprocessor.java ! test/jdk/lambda/separate/ClassToInterfaceConverter.java ! test/jdk/lambda/separate/Compiler.java ! test/jdk/lambda/separate/DirectedClassLoader.java ! test/jdk/lambda/separate/SourceModel.java ! test/jdk/lambda/separate/TestHarness.java ! test/jdk/lambda/vm/DefaultMethodRegressionTests.java ! test/jdk/lambda/vm/DefaultMethodsTest.java ! test/jdk/lambda/vm/InterfaceAccessFlagsTest.java Changeset: 3c08c9ebd1fb Author: katleman Date: 2013-05-30 10:58 -0700 URL: http://hg.openjdk.java.net/hsx/hotspot-gc/jdk/rev/3c08c9ebd1fb Added tag jdk8-b92 for changeset a2a2a91075ad ! .hgtags

1 0

hg: hsx/hotspot-gc/jaxws: Added tag jdk8-b92 for changeset a0f604766ca1
by john.coomes＠oracle.com 31 May '13

31 May '13

Changeset: 7386eca865e1 Author: katleman Date: 2013-05-30 10:58 -0700 URL: http://hg.openjdk.java.net/hsx/hotspot-gc/jaxws/rev/7386eca865e1 Added tag jdk8-b92 for changeset a0f604766ca1 ! .hgtags

1 0

hg: hsx/hotspot-gc/jaxp: 2 new changesets
by john.coomes＠oracle.com 31 May '13

31 May '13

Changeset: 1ab5d8d6eab8 Author: katleman Date: 2013-05-29 10:15 -0700 URL: http://hg.openjdk.java.net/hsx/hotspot-gc/jaxp/rev/1ab5d8d6eab8 8015525: JDK8 b91 source with GPL header errors Reviewed-by: dholmes, lancea ! src/com/sun/org/apache/xerces/internal/impl/XMLEntityManager.java Changeset: a6596a06e10b Author: katleman Date: 2013-05-30 10:58 -0700 URL: http://hg.openjdk.java.net/hsx/hotspot-gc/jaxp/rev/a6596a06e10b Added tag jdk8-b92 for changeset 1ab5d8d6eab8 ! .hgtags

1 0