RFR (XXS): 8048088: Conservative maximum heap alignment should take vm_allocation_granularity into account

Bengt Rutisson bengt.rutisson at oracle.com
Mon Jul 7 12:55:05 UTC 2014


On 2014-07-07 14:47, Thomas Schatzl wrote:
> Hi all,
>
>    can I have reviews for the following small change? It fixes a at this
> time benign issue when calculating the conservative maximum heap
> alignment.
>
> In particular, it does not take the os::vm_allocation_granularity() into
> account.
>
> It is benign for two reasons:
>    - at this time, the conservative maximum heap alignment is always
> larger than the vm allocation granularity (64k on Windows, vm_page_size
> on others), which is 2M for some reason.
>    - even if that were not the case, code further ahead would cause
> assertion failures. I.e. the ReservedHeap checks.
>
> As it is a bug, although it is hidden by other code, and caught already
> later too, I thought it would still be worth fixing this.
>
> This code touches runtime code, so I cc'ed hotspot-runtime-dev.
>
> Webbug:
> http://cr.openjdk.java.net/~tschatzl/8048088/webrev/
>
> CR:
> https://bugs.openjdk.java.net/browse/JDK-8048088

Looks good to me.

Bengt


>
> Testing:
> JPRT
>
> Thanks,
>    Thomas
>
>



More information about the hotspot-gc-dev mailing list