RFR (XS) 8071770 G1 does not implement millis_since_last_gc which is needed by RMI GC

Jon Masamitsu jon.masamitsu at oracle.com
Mon Jul 25 17:50:52 UTC 2016


Joe,

Looks good.  One minor point.

http://cr.openjdk.java.net/~jprovino/8071770/webrev.01/src/share/vm/gc/g1/g1CollectedHeap.cpp.frames.html

I think the comment

2477 // See the notes in GenCollectedHeap::millis_since_last_gc()
2478 // for more information about the implementation.

explains why the  test is needed.

2483 if (ret_val < 0) {
2484 log_warning(gc)("millis_since_last_gc() would return : " JLONG_FORMAT
2485 ". returning zero instead.", ret_val);
2486     return 0;
2487 }


so it might be better if the comment is after the "if"

2483 if (ret_val < 0) {2477 // See the notes in 
GenCollectedHeap::millis_since_last_gc()
2478      // for more information about the implementation.

You can decide.  I don't need another webrev.

Jon


On 07/25/2016 08:48 AM, Joseph Provino wrote:
>
> I just need one more review.  ;-)
>
>
> http://cr.openjdk.java.net/~jprovino/8071770/webrev.01
>
> JDK-8071770 <https://bugs.openjdk.java.net/browse/JDK-8071770> G1 does 
> not implement millis_since_last_gc which is needed by RMI GC
>
> joe
>
>
> On 7/25/2016 10:55 AM, Thomas Schatzl wrote:
>> Hi Joe,
>>
>> On Mon, 2016-07-25 at 10:49 -0400, Joseph Provino wrote:
>>> Can I remove the space after the word "be"?
>>>
>>> Do you want to see another webrev?
>>>
>>> thanks.
>>>
>>> joe
>>>
>>>       log_warning(gc)("Detected clock going backwards. "
>>>         "Milliseconds since last GC would be " JLONG_FORMAT
>>>         ". returning zero instead.", ret_val);
>>>
>>>       log_warning(gc)("Detected clock going backwards. "
>>>         "Milliseconds since last GC would be " JLONG_FORMAT
>>>         ". returning zero instead.", retVal);
>>>
>>   No. Ship it. :)
>>
>> Thomas
>>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/hotspot-gc-dev/attachments/20160725/b7e7f43b/attachment-0001.html>


More information about the hotspot-gc-dev mailing list