RFR (S) 8017243 : 8001345 is incomplete

[Christian Thalinger]

Looks good.  -- Chris

On Jun 24, 2013, at 12:26 PM, Vladimir Kozlov <vladimir.kozlov at oracle.com> wrote:

> Perfect.
> This time lets wait other reviews/comments.
> 
> Thanks,
> Vladimir
> 
> On 6/24/13 12:20 PM, Niclas Adlertz wrote:
>> Thanks for catching that Vladimir.
>> 
>> webrev: http://cr.openjdk.java.net/~adlertz/JDK-8017243/webrev01
>> 
>> Kind Regards,
>> Niclas Adlertz
>> 
>> 
>> On 24 jun 2013, at 21:07, Vladimir Kozlov <Vladimir.Kozlov at oracle.com> wrote:
>> 
>>> Spoke too soon. Please, remove hash_delete() in old 'else' code as Roland pointed before.
>>> 
>>> thanks,
>>> Vladimir
>>> 
>>> On 6/24/13 12:01 PM, Vladimir Kozlov wrote:
>>>> It is good.
>>>> 
>>>> Thanks,
>>>> Vladimir
>>>> 
>>>> On 6/24/13 4:08 AM, Niclas Adlertz wrote:
>>>>> Hi all.
>>>>> 
>>>>> "With compressed oops membar points to Load node through DecodeN node
>>>>> (as in bug 8001345 case). When membar is the only user of DecodeN node
>>>>> it does not mean the Load node has only one user. There is
>>>>> optimization which skip EncodeP->DecodeN chain so Load can have other
>>>>> uses.
>>>>> When membar points to DecodeN and only user, DecodeN node should be
>>>>> replaced by referenced Load node if it has more then one user."
>>>>> 
>>>>> WEBREV: http://cr.openjdk.java.net/~adlertz/JDK-8017243/webrev00/
>>>>> JBS: https://jbs.oracle.com/bugs/browse/JDK-8017243
>>>>> 
>>>>> Kind Regards,
>>>>> Niclas Adlertz
>>>>> 
>>>>> 
>>