RFR: 8135018: AARCH64: Missing memory barriers for CMS collector

Dean Long dean.long at oracle.com
Thu Sep 3 20:18:23 UTC 2015

I forgot to ask, does this fix a real problem seen during testing? If 
so, what's the easiest way to reproduce the problem?


On 9/3/2015 11:30 AM, Andrew Haley wrote:
> On 09/03/2015 06:52 PM, Dean Long wrote:
>> I think you should include the GC alias too.  I thought earlier
>> discussions decided that a barrier wasn't needed.
> I don't remember exactly that.  There was some talk about alternative
> algorithms, but I don't think it's come to anything yet.  There was a
> long discussion about conditional marking, and the StoreLoad barrier
> went in.
>> In LIRGenerator::do_StoreIndexed, it looks like the barrier is being
>> done twice, first in store_check, and again in post_barrier. Doing
>> it in post_barrier makes more sense to me than doing it in
>> store_check.
> Okay, I'll check.
> Thanks,
> Andrew.

More information about the hotspot-compiler-dev mailing list