RFR(S): 8028514: PPC64: Fix C++ Interpreter after '7195622: CheckUnhandledOops has limited usefulness now'

[Vladimir Kozlov]

I submitted JPRT job to push the fix.

Thanks,
Vladimir

On 11/19/13 11:11 AM, David Holmes wrote:
> Thanks Volker - nothing further from me.
>
> David
>
> On 19/11/2013 8:08 PM, Volker Simonis wrote:
>> Hi Lois,
>>
>> thanks for your review. Here's the new webrev:
>>
>> http://cr.openjdk.java.net/~simonis/webrevs/8028514.v2/
>>
>> Please find my comments inline:
>>
>> On Tue, Nov 19, 2013 at 3:35 AM, Lois Foltan <lois.foltan at oracle.com>
>> wrote:
>>> Hi Volker,
>>>
>>> I do have a few review comments:
>>>
>>>    src/cpu/ppc/vm/frame_ppc.cpp
>>>      - The "cast_to_oop<void *>(NULL)" I believe could be simplified to
>>> "(oop)NULL" see similar code in
>>>         src/cpu/x86/vm/frame_x86.cpp.
>>>
>>
>> I was not sure here and wanted to make the conversion explicit,
>> because NULL is actually the integer constant 'zero' and there's no
>> special oop constructor for integers. But if the compiler doesn't
>> complain and if it works on x86 I can change it as you suggested.
>>
>>>    src/share/vm/interpreter/bytecodeInterpreter.cpp
>>>      - I have attached the recent RFR for supporting
>>> CHECK_UNHANDLED_OOPS
>>> across the platforms.
>>>        It contains the summary outlining the changes.  Please change
>>> your
>>> "cast_from_oop<uintptr_t>"
>>>        within tty->print_cr() to "(void *)" as described in summary
>>> note item
>>> #3.
>>
>> You're right, I missed the (void*)operator in the oop class. I changed
>> it the way you suggested.
>>
>> Thanks,
>> Volker
>>
>>>
>>> Thank you,
>>> Lois
>>>
>>>
>>> On 11/18/2013 11:45 AM, Volker Simonis wrote:
>>>>
>>>> Hi,
>>>>
>>>> could you please review the following small webrev which fixes the
>>>> CPP-interpreter after "CheckUnhandledOops" was re-enabled in the
>>>> fastdebug build:
>>>>
>>>> http://cr.openjdk.java.net/~simonis/webrevs/8028514/
>>>>
>>>> All the changes are in CPP-interpreter specific files or in code which
>>>> is protected by "#ifdef  CC_INTERP". Most changes just switch
>>>> old-style "(oop)"casts to new-style "cast_to_oop()" conversions.
>>>>
>>>> Thank you and best regards,
>>>> Volker
>>>
>>>