RFR: 8027230: Overflow in java.lang.instrument.Instrumentation.getObjectSize() method
David Holmes
david.holmes at oracle.com
Wed May 21 05:54:04 UTC 2014
On 21/05/2014 1:19 AM, Leonid Mesnik wrote:
> Peter
>
> 35 * @run main/othervm -Xmx4000m -javaagent:agent.jar
> GetObjectSizeOverflowAgent
>
> I think that "-Xmx4000m" cause test failure for 32-bit VM. It would be
> better to use another 1 process builder.
If you need a 4GB heap to test this you will have to limit it to 64-bit
platforms.
> Also I think you need to add your test into need_jdk because you use
> jar. Could you please check this with embedded team.
JDKToolFinder should get jar from the compile JDK rather than the test
JDK so that should be okay.
However the use of the agent/instrumentation is limited to compact3
profile (if java.lang.instrument is used) and not the minimal VM, so
changes are needed in TEST.groups. It looks like we have a number of
missing updates to the groups file for the test/serviceability/jvmti tests.
David
-----
> Leonid
>
> On 20.05.2014 19:02, Peter Allwin wrote:
>> Hello!
>>
>> Please review this simple fix for an integer overflow in JVMTI
>> GetObjectSize().
>>
>> webrev: http://cr.openjdk.java.net/~allwin/8027230/webrev.00/
>> cr: https://bugs.openjdk.java.net/browse/JDK-8027230
>>
>>
>> Testing:
>> New regression test
>> nsk.quick-jvmti.testlist
>>
>> Thanks!
>> /peter
>
More information about the serviceability-dev
mailing list