RFR (S): 8074546: Rename and clean up the ParGCAllocBuffer class

Joseph Provino joseph.provino at oracle.com
Tue Apr 7 16:16:22 UTC 2015


I made the changes recommended by Kim.

It passes jprt.

As Kim mentioned, removing plab.h from g1CollectedHeap.hpp could 
potentially show
places where plab.h should be included but isn't.

Is there something else I should do to verify there isn't a problem?

joe

On 4/7/2015 4:31 AM, Stefan Karlsson wrote:
>
> On 2015-04-03 20:22, Kim Barrett wrote:
>> On Apr 3, 2015, at 1:32 PM, Kim Barrett <kim.barrett at oracle.com> wrote:
>>> On Apr 3, 2015, at 11:30 AM, Joseph Provino 
>>> <joseph.provino at oracle.com> wrote:
>>>> ParGCAllocBuffer is changed to PLAB and G1ParGCAllocBuffer is 
>>>> changed to G1PLAB.
>>>>
>>>> parGCAllocBuffer.* is changed to plab*.
>>>>
>>>> CR:  https://bugs.openjdk.java.net/browse/JDK-8074546
>>>>
>>>> webrev: http://cr.openjdk.java.net/~jprovino/8074546/webrev.00
>>>>
>>>> Test:  jprt
>>> Looks good.
>> Foo!  I got interrupted and forgot I had a couple of minor comments.
>>
>> ------------------------------------------------------------------------------ 
>>
>> src/share/vm/gc_implementation/shared/plab.hpp
>>    25 #ifndef SHARE_VM_GC_IMPLEMENTATION_PARNEW_PLAB_HPP
>>    26 #define SHARE_VM_GC_IMPLEMENTATION_PARNEW_PLAB_HPP
>> ...
>>   197 #endif // SHARE_VM_GC_IMPLEMENTATION_PARNEW_PLAB_HPP
>>
>> PARNEW => SHARED
>>
>> [Apparently that didn't get updated when the file was moved from 
>> /parNew/
>> to /shared/ several years ago.]
>>
>> ------------------------------------------------------------------------------ 
>>
>> src/share/vm/gc_implementation/g1/g1CollectedHeap.hpp
>>    42 #include "gc_implementation/shared/plab.hpp"
>>
>> Since there are no other changes in this file, I think this include
>> isn't needed at all by this file.  Of course, taking it out might
>> uncover places that should be including it but aren't...
>>
>>
>
> Looks good. It would be great if you could fix the nits Kim mentions.
>
> Thanks,
> StefanK



More information about the hotspot-gc-dev mailing list