RFR (S) 8031818: Experimental VM flag for enforcing safe object construction

[Vladimir Kozlov]

Aleksey,

hs-comp is closed for almost 2 weeks until SQE done all testing. Only 
ppc merge related fixes are allowed.

You check is_initializer() for all types of fields in do_exits(). But 
is_final() is also set for @Stable field which could any methods based 
on the comment in do_put_xxx().

Thanks,
Vladimir

On 1/30/14 12:15 PM, Aleksey Shipilev wrote:
> On 01/22/2014 11:13 PM, Aleksey Shipilev wrote:
>> On 01/22/2014 11:07 PM, Vladimir Kozlov wrote:
>>> Question first: can you wait about 2 weeks when we merge ppc64 changes?
>>
>> I think we can wait. I don't want to collide with PPC merge either. In
>> fact, it would be even more convenient for us to grab PPC C2 in the
>> experiments from the mainline.
>
> Are we open already? I see PPC changes had collided with my patch. Even
> if it's too early to push the patch, I merged the upstream changes into
> the updated webrev:
>    http://cr.openjdk.java.net/~shade/8031818/webrev.01/
>
> Instead of introducing a new method, I rewired the predicate and
> rearranged the comments to make its structure clear. I think extracting
> method is not required now. Also, a few typos corrected.
>
> Testing:
>   - full cycle JPRT
>
> -Aleksey.
>