RFR (S) 7060111 : race condition in VMError::report_and_die()

[David Holmes]

Hi Fred,

Sorry I didn't get to this before you pushed.

Someone closed 8016465 as a duplicate of this bug, but it isn't actually 
a dup as I indicated in the comments. As you have not fixed the issue 
reported in 8016465 that bug will need to be re-opened and fixed separately.

David
-----

On 27/06/2013 10:30 PM, frederic parain wrote:
> Please review this small fix:
>
> http://cr.openjdk.java.net/~fparain/7060111/webrev.00/
>
> Bug: 7060111 : race condition in VMError::report_and_die()
>
> https://jbs.oracle.com/bugs/browse/JDK-7060111
>
> The bug is well described in its report. This fix uses
> class static fields initialized at library load time
> instead of static local variables lazily initialized.
>
> Test:
>
> The race condition is so tricky (2 threads must invoke
> VMError::report_and_die() almost at the same time), that I
> didn't try to produce a unit test for it. But the fix is
> so simple that it doesn't look risky.
>
> Thanks,
>
> Fred
>