RFR(S): 7200261: G1: Liveness counting inconsistencies during marking verification

Jon Masamitsu jon.masamitsu at oracle.com
Tue Sep 25 13:41:21 UTC 2012


Is it common practice in G1 code to have the ranges
be inclusive?


PS.  Sorry for being tardy on my reply - so much mail,
so little time :-)

On 09/24/12 10:03, John Cuthbertson wrote:
> Hi Jon,
> Thanks for the comments. I'm going to refer you to my reply to Jesper. 
> If people feel strongly about making the routine exclusive - I'll make 
> it so.
> JohnC
> On 09/21/12 22:48, Jon Masamitsu wrote:
>> I'm used to seeing a range like [start, end).  That the second index 
>> is named
>> last_idx helps but if I were just looking at the call site, I would 
>> have guessed
>> wrongly - i.e., thinking it was [start, end).
>> Jon
>> On 9/21/2012 9:16 PM, Jesper Wilhelmsson wrote:
>>> John,
>>> Looks good!
>>> Would it make sense to change set_card_bitmap_range to be exclusive, 
>>> or even to take a start and a size of the area? I think inclusive 
>>> functions like this one are unintuitive, especially when the only 
>>> use of last_idx is used with a +1. Maybe that's a different change?
>>> /Jesper
>>> 22 sep 2012 kl. 01:37 skrev John 
>>> Cuthbertson<john.cuthbertson at oracle.com>:
>>>> Hi Everyone,
>>>> Can I have a couple of volunteers look over the fix for this CR? 
>>>> The webrev can be found at: 
>>>> http://cr.openjdk.java.net/~johnc/7200261/webrev.0/
>>>> Summary:
>>>> The clipping in the code that sets the bits for a range of cards in 
>>>> the "expected" card bitmap that we check the liveness accounting 
>>>> data against was incorrect. This could lead to spurious 
>>>> verification failures. In addition to fixing the clipping, I've 
>>>> upleveled this routine and moved it into ConcurrentMark and now use 
>>>> it to generate the real liveness data.
>>>> Testing:
>>>> The failing test cases with marking verification; jprt.
>>>> Thanks,
>>>> JohnC

More information about the hotspot-gc-dev mailing list