RFR (M): CR 8003985: Support @Contended Annotation - JEP 142
serguei.spitsyn at oracle.com
serguei.spitsyn at oracle.com
Thu Nov 29 12:14:46 PST 2012
I'm not pretending to be a reviewer yet.
Just a small comment.
*/src/share/vm/oops/fieldInfo.hpp:*
46 // as an array of 8 shorts
47 enum FieldOffset {
48 access_flags_offset = 0,
49 name_index_offset = 1,
50 signature_index_offset = 2,
51 initval_index_offset = 3,
52 low_offset = 4,
53 high_offset = 5,
54 is_contended_offset = 6,
55 contended_group_offset = 7,
56 field_slots = 8
57 };
It seems the comment about array of 8 shorts is incorrect. Should it be 9?
The comment before the fix was about 7 shorts and you add two new elements:
46 // as an array of 7 shorts
47 enum FieldOffset {
48 access_flags_offset = 0,
49 name_index_offset = 1,
50 signature_index_offset = 2,
51 initval_index_offset = 3,
52 low_offset = 4,
53 high_offset = 5,
54 field_slots = 6
55 };
Thanks,
Serguei
On 11/29/12 2:52 AM, Aleksey Shipilev wrote:
> Anyone?
>
> Thanks,
> Aleksey.
>
> On 11/27/2012 09:57 PM, Aleksey Shipilev wrote:
>> Hi,
>>
>> This is the daizy-fresh thread for reviewing the actual changes to
>> support the JEP-142 in HotSpot. Please address API issues about
>> @Contended to concurrency-interest list, Doug had started the thread here:
>>
>> http://cs.oswego.edu/pipermail/concurrency-interest/2012-November/010208.html
>>
>> We are not discussing the design, API and all the general stuff here,
>> only the reference implementation. I would like to have someone extra
>> knowledgeable in HotSpot codebase to see whether I do anything stupid there.
>>
>> The webrev for the change is here:
>> http://shipilev.net/pub/jdk/hotspot/contended/webrev-4/
>>
>> Notable differences against latest version:
>> - a little denser packing in some corner cases
>> - support @Contended for static fields
>> - moved @Contended to sun.misc (go and nag at c-i@ about this)
>> - bugfixes
>>
>> Testing:
>> - hotspot/test/runtime/8003985/Test8003985 on Linux x86_64
>> - jprt, specjvm98 quick run
>> - jprt full cycle is underway, looking good
>>
>> Thanks,
>> Aleksey.
>>
More information about the hotspot-dev
mailing list