RFR(S): Implementation of aarch64 interpreter matrix barrier

Andrew Haley aph at redhat.com
Wed Jun 21 16:32:18 UTC 2017


On 21/06/17 17:15, Aleksey Shipilev wrote:
> On 06/21/2017 04:51 PM, Zhengyu Gu wrote:
>> Thanks for the review, Andrew and Aleksey,
>>
>> I updated patch according to your comments:
>>
>> Webrev:
>> http://cr.openjdk.java.net/~zgu/shenandoah/aarch64_int_matrix_barrier/webrev.01/
> 
> *) This line smells odd:
> 
> 3880   unsigned long offset;
> 3881   adrp(rscratch2, ExternalAddress(ShenandoahHeap::heap()->base()), offset);
> 
> ...is that ExternalAddress(...base(), (unsigned long)0)?

No, it's correct.  Offset is passed by reference to adrp().  Sorry
about that, but it seemed like a good idea at the time...  :-)

-- 
Andrew Haley
Java Platform Lead Engineer
Red Hat UK Ltd. <https://www.redhat.com>
EAC8 43EB D3EF DB98 CC77 2FAD A5CD 6035 332F A671


More information about the shenandoah-dev mailing list