RFR: JDK-8213615: GC/C2 abstraction for escape analysis

Vladimir Kozlov vladimir.kozlov at oracle.com
Wed Nov 14 18:56:37 UTC 2018


G1 and ZGC code refactoring matches previous code.
add_to_congraph_unsafe_access() is also good.

Why add_final_edges_unsafe_access() does not include 'if (opcode == Op_GetAndSetP' code? You broke sequence where 
(adr_type == NULL) should be checked first.

Also fail check is wrong. Should be:

        if (add_final_edges_unsafe_access(n, opcode)) {
          break;
        }
        ELSE_FAIL("Op_StoreP");

Why you moved record_for_optimizer() to .cpp file?

Thanks,
Vladimir

On 11/13/18 9:05 AM, Roman Kennke wrote:
> Sure no problem. Thank you!
> 
> Roman
> 
> 
>> I have to review it in details. Give me some time.
>>
>> Thanks,
>> Vladimir
>>
>> On 11/13/18 4:24 AM, Roman Kennke wrote:
>>> Thanks Roland for the review!
>>>
>>> Vladimir: do you want to take a deeper look too? Or can I consider your
>>> 'I like the proposal from JIT POV' as 'reviewed' ? ;-)
>>>
>>> Roman
>>>
>>>>> http://cr.openjdk.java.net/~rkennke/JDK-8213615/webrev.01/
>>>>
>>>> Looks ok to me.
>>>>
>>>> Roland.
>>>>
>>>
> 


More information about the hotspot-gc-dev mailing list