Pls (re)review 6173675/7003271: per-thread memory allocation measurement

[Alan Bateman]

Mandy Chung wrote:
>  On 1/9/11 5:10 PM, Paul Hohensee wrote:
>> Thanks to all who reviewed the previous webrev.  Revised versions are 
>> here
>>
>> http://cr.openjdk.java.net/~phh/6173675/webrev.02/
>>
> Looks good.  Thanks for updating it and adding the tests.
>
> As for the tests, the existing java/lang/management regression tests 
> definitely need upgrade and improvement.  They were written before 
> java.util.concurrent locks were added in 1.5 and also the tests have 
> run into some timing issue between the work done by threads and the 
> test verification.  I'm fine for pushing the tests you have and then 
> follow up with a CR as David suggests.
>
> Thanks
> Mandy
One thing on the tests, what would you think about disabling the per CPU 
time and memory allocation in a finally block so that they do clean-up 
in the event of a test failure. It would be useful once we get to run 
these tests in samevm mode (I think currently the Makefile runs these 
tests in othervm mode so it's not a problem if the tests don't clean-up).

-Alan