RFR 8153302 [BACKOUT] STW phases at Concurrent GC should count in PerfCounter

Daniel D. Daugherty daniel.daugherty at oracle.com
Fri Apr 1 19:11:18 UTC 2016


On 4/1/16 1:07 PM, Jon Masamitsu wrote:
>
>
> On 4/1/2016 11:41 AM, Daniel D. Daugherty wrote:
>> Greetings,
>>
>> It appears that the fix for this bug:
>>
>> JDK-8151674 STW phases at Concurrent GC should count in PerfCounter
>> https://bugs.openjdk.java.net/browse/JDK-8151674
>>
>> has caused regressions in the 2016.03.31 JDK9-hs-rt nightly:
>>
>> JDK-8153280 sun/tools/jstatd/TestJstatd* tests fail with 
>> java.io.IOException:
>>             Stream closed and missing headline
>> https://bugs.openjdk.java.net/browse/JDK-8153280
>>
>> JDK-8153283 Multiple jstat failures in hs nightly
>> https://bugs.openjdk.java.net/browse/JDK-8153283
>>
>>
>> I've prepared backout changesets for the jdk and hotspot repos and
>> I'm running the affected testsuites to verify that the regression
>> goes away.
>>
>> Here are the webrev URLs:
>>
>> http://cr.openjdk.java.net/~dcubed/8153302-webrev/0-jdk9-hs-rt-jdk/
>>
>> http://cr.openjdk.java.net/~dcubed/8153302-webrev/0-jdk9-hs-rt-hotspot/
>
> Changes look correct.  Thanks.

Thanks for the fast review Jon!

Dan


>
> Jon
>
>>
>> Here are the changeset links for JDK-8151674:
>>
>> http://hg.openjdk.java.net/jdk9/hs-rt/jdk/rev/8fa63e27e764
>>
>> http://hg.openjdk.java.net/jdk9/hs-rt/hotspot/rev/252b571bbb86
>>
>> Thanks, in advance, for a couple of quick reviews.
>>
>> Gory details below...
>>
>> Dan
>>
>>
>>
>>
>> Here's my jdk repo sanity check note:
>>
>> Daniel Daugherty added a comment - 25 minutes ago
>> Sanity checking the jdk repo portion of the backout:
>>
>> $ hg log -r tip
>> changeset: 13957:8fa63e27e764
>> tag: tip
>> user: ysuenaga
>> date: Wed Mar 30 21:05:35 2016 +0900
>> summary: 8151674: STW phases at Concurrent GC should count in 
>> PerfCounter
>>
>> $ hg diff -r 13956
>>
>> The changeset for 8151674 is the tip in the jdk repo so the backout 
>> is easy.
>>
>>
>> Here's my hotspot repo sanity check note:
>>
>> Daniel Daugherty added a comment - 22 minutes ago
>> Sanity checking the hotspot repo portion of the backout:
>>
>> $ hg log -r tip
>> changeset: 10684:d9d1edafdf2d
>> tag: tip
>> user: mchernov
>> date: Fri Apr 01 16:15:37 2016 +0300
>> summary: 8148397: Create new tests for IHOP
>>
>> $ hg log -r 252b571bbb86
>> changeset: 10670:252b571bbb86
>> parent: 10665:500776d80c5b
>> user: ysuenaga
>> date: Wed Mar 30 21:05:13 2016 +0900
>> summary: 8151674: STW phases at Concurrent GC should count in 
>> PerfCounter
>>
>>
>> A total of 14 changeset were pushed after the changeset for 8151674.
>>
>> $ cat files.list
>> src/share/vm/gc/cms/concurrentMarkSweepGeneration.cpp
>> src/share/vm/gc/cms/concurrentMarkSweepGeneration.hpp
>> src/share/vm/gc/g1/g1ConcurrentMark.cpp
>> src/share/vm/gc/g1/g1MonitoringSupport.cpp
>> src/share/vm/gc/g1/g1MonitoringSupport.hpp
>> src/share/vm/gc/g1/vm_operations_g1.cpp
>> src/share/vm/gc/shared/vmGCOperations.hpp
>>
>> $ hg diff -r 10665 `cat files.list`
>>
>> Fortunately, there were no overlaps in the files changed by 8151674
>> so the backout does not require any merging.
>



More information about the hotspot-gc-dev mailing list