RFR: 8036696: Add metaspace gc threshold to metaspace summary trace event

[Jon Masamitsu]

On 3/10/14 2:22 AM, Erik Helin wrote:
> Hi Jon,
>
> thanks for having a look at the patch!
>
> On 2014-03-06 23:30, Jon Masamitsu wrote:
>> Erik,
>>
>> Why the choice of the variable name _gc_threshold instead of the
>> variable name _capacity_until_gc (the same variable name as
>> in Metaspace)?  Using the label "GC Threshold" seems ok.  I
>> can understand not making that title too geeky.
>
> Since the the meaning of the word "capacity" has caused me (us?) some 
> problems in the metaspace code (is it words/bytes? is it 
> committed/used or something else?), I believe it is better to use that 
> does not imply to "what" the high-water mark is compared. The name 
> _capacity_until_gc implies that we compare the value to some kind of 
> capacity, whereas the name gc_threshold just states that this is a 
> some kind of limit.
>
> Ideally, I would like to see us rename MetaspaceGC::_capacity_until_gc 
> to MetaspaceGC::_gc_threshold, but that is another patch.
>
> What do you think?

Can you correct _capactiy_until_gc to a more  correct and precise name?
Or at least change _gc_threshold to a more correct and precise name
and make _capacity_until_gc follow later.

_gc_threshold seems to me has all the ambiguities of _capacity_until_gc.

Jon

>
> Thanks,
> Erik
>
>> Otherwise, looks good.
>>
>> Jon
>>
>> On 3/6/2014 12:07 AM, Erik Helin wrote:
>>> Hi all,
>>>
>>> this change add the metaspace GC threshold
>>> (Metaspace::_capacity_until_gc) to the trace event
>>> vm/gc/heap/metaspace_summary.
>>>
>>> Enhancement:
>>> https://bugs.openjdk.java.net/browse/JDK-8036696
>>>
>>> Webrev:
>>> http://cr.openjdk.java.net/~ehelin/8036696/webrev.00/
>>>
>>> Testing:
>>> - JFR JTREG tests
>>> - JPRT
>>>
>>> Thanks,
>>> Erik
>>