RFR(S): 8040803: G1: Concurrent mark hangs when mark stack overflows

Bengt Rutisson bengt.rutisson at oracle.com
Mon May 5 16:11:29 UTC 2014


Hi Per,

Looks good. 

Bengt

> 5 maj 2014 kl. 15:27 skrev Per Liden <per.liden at oracle.com>:
> 
> Hi Bengt,
> 
>> On 05/02/2014 10:33 AM, Bengt Rutisson wrote:
>> 
>> Hi Per,
>> 
>>> On 2014-04-30 15:04, Per Liden wrote:
>>> Hi,
>>> 
>>> Could I please have a couple of reviews in this bug fix:
>>> 
>>> Summary: G1's concurrent marking can potentially hang forever if the
>>> global mark stack overflows and immediately after that a Full GC
>>> happens, which tries to abort the marking. The reason is that there's
>>> a race between detecting the overflow situation and detecting the
>>> abort signal. Threads detecting the overflow situation first will go
>>> into the overflow protocol and wait on a barrier for all threads to
>>> reach this state. However, threads detecting the abort signal first
>>> will terminate and never participate in the barrier.
>>> 
>>> This patch introduces an abort state and function on the
>>> WorkGangBarrierSync class, to unblock any threads waiting for the
>>> barrier to complete when the concurrent mark is aborted.
>>> 
>>> Bug: https://bugs.openjdk.java.net/browse/JDK-8040803
>>> Webrev: http://cr.openjdk.java.net/~pliden/8040803/webrev.0/
>> 
>> The change looks good.
> 
> Thanks for looking at the patch Bengt.
> 
>> 
>> One minor improvement would be to try to use an early exit in
>> ConcurrentMark::enter_first_sync_barrier() when the barrier fails.
> 
> Here's an updated webrev with early exit:
> 
> http://cr.openjdk.java.net/~pliden/8040803/webrev.1/
> 
> /Per
> 
>> 
>> Thanks,
>> Bengt
>> 
>>> 
>>> /Per
> 


More information about the hotspot-gc-dev mailing list