Differences between openjdk6 hotspot and hs14+

[Joseph D. Darcy]

Martin Buchholz wrote:
> Looks good.
>
> Martin
>
> On Mon, Nov 23, 2009 at 12:16, Andrew John Hughes
> <gnu_andrew at member.fsf.org> wrote:
>   
>> 2009/11/11 Martin Buchholz <martinrb at google.com>:
>>     
>>> Hi Andrew,
>>>
>>> I'm checking the current state of hotspot 14 in openjdk6,
>>> and noticed some discrepancies I can't explain.
>>>
>>> This hunk:
>>>
>>> --- /home/martinrb/ws/hsx14-master/src/share/vm/runtime/globals.hpp
>>> 2009-05-13 14:49:39.000000000 -0700
>>> +++ /home/martinrb/ws/openjdk6/hotspot/src/share/vm/runtime/globals.hpp
>>> 2009-11-04 15:34:13.000000000 -0800
>>> @@ -1494,7 +1497,7 @@
>>>            "CMSPrecleanNumerator:CMSPrecleanDenominator yields convergence"
>>> \
>>>            " ratio")
>>> \
>>>
>>> \
>>> -  product(bool, CMSPrecleanRefLists1, true,
>>> \
>>> +  product(bool, CMSPrecleanRefLists1, false,
>>> \
>>>            "Preclean ref lists during (initial) preclean phase")
>>> \
>>>
>>> \
>>>    product(bool, CMSPrecleanRefLists2, false,
>>> \
>>>
>>>
>>> is from
>>>
>>> # HG changeset patch
>>> # User ysr
>>> # Date 1227212861 28800
>>> # Node ID 00b023ae2d78987955a9647b6d6d457bd39dec7e
>>> # Parent  b5e603f2e02480180ace404141897cf8f6266e93
>>> 6722113: CMS: Incorrect overflow handling during precleaning of Reference
>>> lists
>>> Summary: When we encounter marking stack overflow during precleaning of
>>> Reference lists, we were using the overflow list mechanism, which can cause
>>> problems on account of mutating the mark word of the header because of
>>> conflicts with mutator accesses and updates of that field. Instead we should
>>> use the usual mechanism for overflow handling in concurrent phases, namely
>>> dirtying of the card on which the overflowed object lies. Since precleaning
>>> effectively does a form of discovered list processing, albeit with discovery
>>> enabled, we needed to adjust some code to be correct in the face of
>>> interleaved processing and discovery.
>>> Reviewed-by: apetrusenko, jcoomes
>>>
>>>
>>> Probably inadvertently reverted during a merge?
>>>
>>>
>>> I found this by reviewing changes thusly
>>>
>>> diff -Bw -x .hg -u -r  ~/ws/hsx14-master/ ~/ws/openjdk6/hotspot/
>>>
>>> Martin
>>>
>>>       
>> Yes, looks like it was reverted in:
>>
>> http://hg.openjdk.java.net/jdk6/jdk6/hotspot/rev/1ba7ca27e707
>>
>> (the big merge)
>>
>> Here's a webrev for the change:
>>
>> http://cr.openjdk.java.net/~andrew/jdk6-hs14-merge/webrev.05/
>>
>> Joe, can you please approve this and allocate it a bug ID?
>>
>> Thanks,
>> --
>> Andrew :-)
>>
>> Free Java Software Engineer
>> Red Hat, Inc. (http://www.redhat.com)
>>
>> Support Free Java!
>> Contribute to GNU Classpath and the OpenJDK
>> http://www.gnu.org/software/classpath
>> http://openjdk.java.net
>>
>> PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
>> Fingerprint: F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8
>>
>>     
Martin, good catch finding this problem!  Andrew, I approve your fix; 
please you bug 6904255 "Fix HS14 merge issue in OpenJDK 6."

Thanks,

-Joe