RFR(S) 8156156: Add module specific NMT MemoryType

[harold seigel]

Hi Coleen,

Thanks for the review!

Harold


On 5/31/2016 4:26 PM, Coleen Phillimore wrote:
>
> Harold, This change looks good.
>
> If I remember correctly, we used to have to explicitly instantiate 
> Hashtable template parameters in hashtable.cpp.  Maybe we don't need 
> to anymore if we don't have any explicit instantiations for mtCode or 
> had one for mtInternal, so I guess it's not necessary.
>
> Coleen
>
>
> On 5/20/16 11:34 AM, Lois Foltan wrote:
>>
>> On 5/20/2016 10:59 AM, harold seigel wrote:
>>> Thanks Lois.
>>>
>>> Why do you think Hashtable<... mtModule> and HashtableEntry<... 
>>> mtModule> are needed.  There are no such entries for mtCode or 
>>> mtInternal.
>>
>> I think it is best practice to explicitly instantiate the types you 
>> know the JVM is going to use/need.  And there seems to be a 
>> precedence for doing so in hashtable.cpp.
>> Lois
>>
>>>
>>> Thanks, Harold
>>>
>>>
>>> On 5/20/2016 10:49 AM, Lois Foltan wrote:
>>>>
>>>> On 5/19/2016 4:00 PM, harold seigel wrote:
>>>>> Hi Lois,
>>>>>
>>>>> Thanks for the suggestions.  I added the allocations that you 
>>>>> listed below.
>>>>>
>>>>> Please see updated webrev: 
>>>>> http://cr.openjdk.java.net/~hseigel/bug_8156156.1/
>>>>
>>>> Thanks Harold for making this change!  Looks great, I think 
>>>> hashtable.cpp also needs a
>>>>
>>>>     Hashtable<Symbol*, mtModule>
>>>>     HashtableEntry<Symbol*, mtModule>
>>>>
>>>> I don't need to see another webrev.
>>>>
>>>> Lois
>>>>
>>>>>
>>>>> Thanks! Harold
>>>>>
>>>>>
>>>>> On 5/19/2016 10:40 AM, Lois Foltan wrote:
>>>>>>
>>>>>> On 5/19/2016 9:05 AM, harold seigel wrote:
>>>>>>> Hi,
>>>>>>>
>>>>>>> Please review this small change to add a module specific memory 
>>>>>>> type to hotspot native memory tracking (NMT).
>>>>>>>
>>>>>>> Open webrev: http://cr.openjdk.java.net/~hseigel/bug_8156156/
>>>>>>>
>>>>>>> Bug: https://bugs.openjdk.java.net/browse/JDK-8156156
>>>>>>>
>>>>>>> The change was tested with an RBT nightly test run, JCK lang and 
>>>>>>> VM tests, and UTE non-colocated quick tests.
>>>>>>
>>>>>> Hi Harold,
>>>>>>
>>>>>> This looks good.  Should we consider changing the 
>>>>>> fixup_module_field_list and ClassLoader's _xpatch_entries, 
>>>>>> _boot_modules_array and _platform_modules_array as well?
>>>>>>
>>>>>> Thanks,
>>>>>> Lois
>>>>>>
>>>>>>>
>>>>>>> Thanks, Harold
>>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>
>