RFR(L): 8059550: Test task: segment overflow w/ empty others

[Vladimir Kozlov]

Looks good.

Thanks,
Vladimir

On 11/20/14 3:30 AM, Pavel Chistyakov wrote:
> Hi Vladimir,
>
> Please take a look to updated webrev: http://cr.openjdk.java.net/~iignatyev/pchistyakov/8059550/webrev.01
> <http://cr.openjdk.java.net/%7Eiignatyev/pchistyakov/8059550/webrev.01>
>
> I calculate blob size using not overall code cache size but using current heap size.
>
> ---------------
> Thanks,
> Pavel
>
> On 19.11.2014 20:08, Vladimir Kozlov wrote:
>> Tobias should look on this.
>>
>> My only concern is the same of a blob is ReservedCodeCacheSize/128 which could be more then NonNmethod segment (5Mb).
>> I would suggest to set ReservedCodeCacheSize on command lines to avoid it.
>>
>> Thanks,
>> Vladimir
>>
>> On 11/18/14 5:47 AM, Pavel Chistyakov wrote:
>>> http://cr.openjdk.java.net/~iignatyev/pchistyakov/8059550/webrev.00/
>>> 127 lines changed: 125 ins; 0 del; 2 mod;
>>>
>>> Hi all,
>>>
>>> please review the patch which adds new tests for Segmented Code Cache feature.
>>>
>>> Test tries to full code heaps using new WB API function (allocateCodeBlob) added by 8059624 and checks that overflow
>>> handling works properly for every code heap.
>>> Also it checks that compilation is disabled on code cache/heap overflow.
>>>
>>> Added new WB functions:
>>> - getCompilationActivityMode() to get compilation mode value (stop_compilation, run_compilation or shutdown_compilation)
>>> - getCodeBlob to get blob object (instance of sun.hotspot.code.CodeBlob class)
>>>
>>> changes in product code:
>>> - get_compilation_activity_mode() added to compileBroker
>>>
>>> testing: locally
>>> jbs: https://bugs.openjdk.java.net/browse/JDK-8059550
>>>
>>> ------------
>>> Thanks,
>>> Pavel
>