RFR: JDK-8077414 PSPromotionLAB _state is unintialized

Jon Masamitsu jon.masamitsu at oracle.com
Mon Apr 13 17:49:06 UTC 2015



On 4/13/2015 6:54 AM, Eric Caspole wrote:
> Hi Jon,
> I didn't look at it that hard to say. One reason seems to be by 
> toggling the state there are asserts to make sure that it is not 
> flushed twice and check the general state of what is going on.

Eric,

Thanks for  the additional explanation.

Jon

> Eric
>
> On 4/10/2015 2:57 PM, Jon Masamitsu wrote:
>> Eric,
>>
>> A bit of an idle question here.  Do you think that LabState (and 
>> hence _state)
>> is needed?  As opposed to using the values of bottom, top, and end to
>> to infer the state of the plab?  "I don't know" is a good enough answer
>> and then I can go look if I am feeling energetic.
>>
>> Thanks.
>>
>> Jon
>>
>> PS.  Changes look good.
>>
>> On 4/10/2015 6:38 AM, Eric Caspole wrote:
>>> While working on another bug I noticed that PSPromotionLAB field 
>>> _state is uninitialized, and since I was staring at this code I 
>>> wondered if this was part of the problem. I don't think so but I 
>>> would like to initialize it so no one wonders in the future. I 
>>> initialize it to "zero_size" since it is otherwise filled in with 
>>> NULL ptrs.
>>>
>>> Bug:
>>>  https://bugs.openjdk.java.net/browse/JDK-8077414
>>>
>>> Webrev:
>>>  http://cr.openjdk.java.net/~ecaspole/JDK-8077414/00/webrev/
>>>
>>> Passes JPRT.
>>> Thanks,
>>> Eric
>>>
>>
>



More information about the hotspot-gc-dev mailing list