RFR(XS) : 8029070 : memory leak in jmm_SetVMGlobal

David Holmes david.holmes at oracle.com
Sun Nov 24 16:45:18 PST 2013


Hi Igor,

On 25/11/2013 3:14 AM, Igor Ignatyev wrote:
> Hi all,
>
> Please review patch (9 lines changed: 6 ins; 3 del; 0 mod)
>
> Problem:
> CommandLineFlags::ccstrAtPut copies a value and expects that caller will
> free the memory of the old value by itslef, but in jmm_SetVMGlobal, it
> doesn't do.
>
> Fix:
> Free the old value in SetVMGlobal.

This looks fine to me.

> The contracts of ccstrAtPut were moved to h-files.

These are in the wrong place in globals.hpp. The contract is only on the 
AtPut variants not the At variants. Not even certain it is on both AtPut 
variants, but presumably so.

David
-----

> testing: jprt
> jbs: https://bugs.openjdk.java.net/browse/JDK-8029070
> webrev: http://cr.openjdk.java.net/~iignatyev/8029070/webrev.00/
>
> Thanks,
> Igor


More information about the hotspot-runtime-dev mailing list