On 26/02/2014 10:56 PM, Peter Levart wrote:
On 02/25/2014 09:38 PM, Brian Burkhalter wrote:
On Feb 20, 2014, at 1:42 AM, Paul Sandoz wrote:
Not sure the static powerCache field, in the original code, needs to
be volatile either:
1137 private static volatile BigInteger[][] powerCache;
Is there consensus on whether "volatile" is necessary here?
I think it has to be volatile. The powerCache implementation was added
in the following changeset:
Definitely has to be volatile for current code. I don't know what I was
looking at when I said it didn't need to be - I didn't see the code that
grows the cache. <shakes head in dismay>
David
-----
http://hg.openjdk.java.net/jdk9/dev/jdk/rev/7546
...and improved later in the following:
http://hg.openjdk.java.net/jdk9/dev/jdk/rev/7586
It uses a copy-on-write technique to extend the cache with new values
when needed. volatile is mandatory here to safely publish the newly
constructed array-of-arrays and the newly constructed sub-array to other
threads. Without volatile, other threads could see null slots where
BigInteger[] and/or BigInteger objects should be...
Regards, Peter
Thanks,
Brian
