Request for review (XS): 7197906: BlockOffsetArray::power_to_cards_back() needs to handle > 32 bit shifts

[Srinivas Ramakrishna]

Hal, thanks for the catch and the fix!

Bengt, might it be possible to host this as a traditional webrev in the
usual place, to simplify review logistics?

thanks!
-- ramki

On Thu, Sep 13, 2012 at 4:55 PM, Jianhao Mo <mojianhao at gmail.com> wrote:

> Hi all,
>
> Attached please find the new patch, modified according to John's advice.
>
> Thanks Bengt again for the help :)
>
> Regards,
>
> Hal
>
> 2012/9/12 Jianhao Mo <mojianhao at gmail.com>
>
>> Hi all,
>>
>> Could I get a couple of reviews for this small patch, please?
>>
>> 7197906: BlockOffsetArray::power_to_cards_back() needs to handle > 32 bit
>> shifts
>>
>> CR link:
>> http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=7197906
>> Webrev:
>> http://cr.openjdk.java.net/~brutisso/7197906/webrev.01/
>> Contributed-by:
>> Hal Mo <kungu.mjh at taobao.com> from Alibaba Group(with OCA)
>>
>> It may take a while before the CR link is publicly accessible.
>>
>> In blockOffsetTable.hpp, there is an unexpected data truncation in
>> power_to_cards_back(uint i), which may lead to crashing the VM on very
>> large GC heaps.
>> The problem could be reproduced easily on machines, that have enough
>> memory, with:
>> $ java -Xmx135g -XX:+UseConcMarkSweepGC
>>
>> The proposed patch fixes this problem, and builds on all OpenJDK
>> supported platforms.
>>
>> I'd like to thank Bengt Rutisson for the initial review and help
>> preparing the CR and Webrev.
>>
>> Regards,
>> Hal
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/hotspot-gc-dev/attachments/20120913/303a3283/attachment.htm>