RFR (S): 6726963: multi_allocate() call does not CHECK_NULL and causes crash in fastdebug bits

[David Holmes]

Hi,

On 31/05/2013 3:25 AM, Mikhailo Seledtsov wrote:
> Hi,
>
> Please review this one-line fix,

Functional fix is fine. Please update the copyright year as this is the 
policy for runtime.

>    and a test code to reproduce this bug.

Your test does not need:

  * @library /testlibrary

or

import com.oracle.java.testlibrary.*;

No need to generate a new webrev though.

Thanks,
David (Reviewer)

> http://cr.openjdk.java.net/~ctornqvi/webrev/6726963/webrev.00/
>
> Overview:
>
>       the test case attempts to test out-of-memory condition allocating
> very large multi-dimensional array.
>
> JVM can throw OOM exception, but should not crash or except in any other
> way.
>
> Fix:
>
> added a null-check when calling multi_allocate while allocating the
> object in corresponding method
>
> of the reflection class
>
> Testing:
>
>      - Created a JTReg test case to reproduce this bug
>
>      - The test case fails before the fix, and passes after the fix.
>
> Tested on 32-bit Linux build.
>
>      - Ran this test case on all supported platforms using JPRT harness
> - PASS
>
>      - Full JPRT run - PASS
>
>      - UTE run with vm.quick.testlist - NO new failures introduced by
> this change
>
> Thank you,
>
> Misha
>