RFR (XXS): 8134857: Inconsistency in maximum TLAB/PLAB size and humonguous object size

[Dmitry Fazunenko]

Jon,

 > What was the answer to this?  Whether being in the g1 directory means 
the
 > test is only run with g1?

All tests in the g1 directory are included into the ':needs_g1' jtreg group.
This group is currently used to exclude g1 specific tests from execution 
on embedded platform where g1 is disabled.
If one will give a command like:
# jtreg -vmoptions:-XX:+UseSerialGC :hotspot_all
The tests from the g1 folder will be executed.
So, the "@requires g1" is needed here to protect from failures due GC 
conflicts.

Thanks,
Dima



On 03.09.2015 21:15, Jon Masamitsu wrote:
> Dima,
>
> One question below.
>
>
> On 9/3/2015 6:24 AM, Thomas Schatzl wrote:
>> Hi Dima,
>>
>>    thanks for the review.
>>
>> On Thu, 2015-09-03 at 15:54 +0300, Dmitry Fazunenko wrote:
>>> Hi Thomas,
>>>
>>> The test you provided looks good.  I have some minor comments.
>>>
>>> 1) @requires vm.gc == null | vm.gc == "G1"
>> I skipped it because the test is located in the g1 directory already,
>> which tests are only executed with G1. Is this the expected way to
>> specify a g1-only test?
>>
>> I added it though.
>
> What was the answer to this?  Whether being in the g1 directory means the
> test is only run with g1?
>
> Jon
>
>>
>>> 2) Please, use 4-space indent in the "runTest" and "main" methods.
>> Okay, sorry.
>>
>>> 3) Adding  a javadoc comment to the "runTest" method would be nice. It
>>> should tell what is tested.
>>> Something like:
>>>     Starts VM with the given G1 region size and PLAB size. It's 
>>> expected
>>> that VM is successfully started
>>>     only if PLAB size is less than or equal to ...
>> Done.
>>
>> New webrevs:
>> http://cr.openjdk.java.net/~tschatzl/8134857/webrev.1/ (full)
>> http://cr.openjdk.java.net/~tschatzl/8134857/webrev.0_to_1/ (diff)
>>
>> Thanks,
>>    Thomas
>>
>>
>