RFR (L) 8154580: Save mirror in interpreter frame to enable cleanups of CLDClosure

[Coleen Phillimore]

On 4/19/16 6:41 PM, Dean Long wrote:
> Looks OK, except I didn't understand the reason for the change to 
> InitArrayShortSize in globals_zero.hpp.

Thank you Dean.  That change is something else I'm sponsoring which I 
needed to get Zero to compile.  It won't be in this changeset.

thanks,
Coleen

>
> dl
>
> On 4/19/2016 12:50 PM, Coleen Phillimore wrote:
>> Summary: GC walks the mirror using OopClosure rather than using 
>> CLDClosure in oops_interpreted_do()
>>
>> See bug for more description and justification.  The changes are 
>> large but very redundant.  The main change is in 
>> TemplateInterpreterGenerator::generate_fixed_frame().
>>
>> open webrev at http://cr.openjdk.java.net/~coleenp/8154580.01/webrev
>> bug link https://bugs.openjdk.java.net/browse/JDK-8154580
>>
>> Tested with hotspot-runtime-nightly and gc-nightly tests.
>>
>> Need testing with ppc and aarch64 open code.  I implemented the 
>> changes but I can't test them.
>>
>> Thanks,
>> Coleen
>