RFR(S): 8134468: Lucene test failures with 32 bit JDK 9b78, Server compiler

Vladimir Kozlov vladimir.kozlov at oracle.com
Wed Sep 2 16:20:22 UTC 2015


Good.

Thanks,
Vladimir

On 9/2/15 1:33 AM, Roland Westrelin wrote:
> http://cr.openjdk.java.net/~roland/8134468/webrev.00/
>
> The computations in ArrayCopyNode::modifies() can overflow on 32bits and the return value can then be incorrect (an array load can then bypass an arraycopy that modifies that array element). Actually I’m not sure why I used intptr_t in the initial change. I probably wanted a 64 bit integer and used the wrong type.
>
> Roland.
>


More information about the hotspot-compiler-dev mailing list