RFR: 6900441 PlatformEvent.park(millis) on Linux could still be affected by changes to the time-of-day clock

David Holmes david.holmes at oracle.com
Mon Sep 16 22:31:52 PDT 2013

This fix has been pushed to hs25 for JDK 8.

For the record testing involved some specific time-jump tests attached 
to the bug report (and further variants to test parkNanos and 
parkUntil); the java/util/concurrent jtreg tests; the jcstress 
concurrency stress tests and the regular JPRT tests.


On 14/09/2013 9:42 AM, David Holmes wrote:
> On 14/09/2013 3:49 AM, Aleksey Shipilev wrote:
>> On 09/13/2013 09:03 PM, Daniel D. Daugherty wrote:
>>>> http://cr.openjdk.java.net/~dholmes/6900441/webrev.v4/
>>> I think you want to get some other eyes on this change. This is
>>> my fourth review and I'm not catching everything. Or I'm catching
>>> something, but not for exactly the right reasons.
>> (Granted, I have mediocre experience with Linux APIs).
>> Nothing pops out for me. Thumbs up.
> Thanks Aleksey.
>> Stylistic:
>>   "++absTime->tv_sec;" bothers me. There seems to be no reason to not to
>> use post-increment rather than let the readers infer what is being
>> incremented, and if the precedence rules play dirty here.
> Noted for future cleanup :)
> David
>> -Aleksey.

More information about the hotspot-dev mailing list