RFR: 8371450: AES performance improvements for key schedule generation

[Shawn M Emery]

On Fri, 7 Nov 2025 17:25:51 GMT, Jamil Nimeh <jnimeh at openjdk.org> wrote:

>> This fix improves performance in the AES key schedule generation by eliminating an unnecessary object and unnecessary mask in the inverse key schedule.
>> 
>> The micro:org.openjdk.bench.javax.crypto.AESReinit benchmark results are improved by 6.96% for arm64 and 7.79% for x86_64.
>> 
>> Thank you @jnimeh for catching the unnecessary byte mask!
>
> src/java.base/share/classes/com/sun/crypto/provider/AES_Crypt.java line 1011:
> 
>> 1009:      */
>> 1010:     private static int subWord(int word) {
>> 1011:         byte b0 = (byte) (word >>> 24);
> 
> Nits (unrelated to this block of code, but I can't put comments on the lines directly since they're not part of the modified code):
> 
> - Lines 37-43: I think you could do an `@link` annotation with <a href></a> surrounding the links.
> - Lines 1004-1005: Looks like a bit of comment rot, I think you just need an `@param` for `word`.

Re: `@link` - I'm not for sure I understand the context and couldn't find an example of this in the existing code base.
Re: `@param` - Good catch.  I've made this update and a couple of others that needed cleaning up.

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/28188#discussion_r2505914134