RFR: 8371450: AES performance improvements for key schedule generation [v3]

[Ferenc Rakoczi]

On Fri, 7 Nov 2025 22:45:33 GMT, Shawn M Emery <duke at openjdk.org> wrote:

>> src/java.base/share/classes/com/sun/crypto/provider/AES_Crypt.java line 1017:
>> 
>>> 1015:                 | ((SBOX[(b1 & 0xF0) >> 4][b1 & 0x0F] & 0xFF) << 16)
>>> 1016:                 | ((SBOX[(b2 & 0xF0) >> 4][b2 & 0x0F] & 0xFF) << 8)
>>> 1017:                 | (SBOX[(word & 0xF0) >> 4][word & 0x0F] & 0xFF);
>> 
>> I think there is no need for any of these bytes. Every index can be computed as "(word >> offset) & 0x0F". Actually, if you define SBOX as a 1-dim array, you can index into it with "(word >> offset) & 0xFF".
>
> Thank you for your review.  The byte assignments were to avoid three redundant shift operations.

What were those "redundant shifts"? 
What I am suggesting is:
 return (SBOX(word >>>24] << 24) |
             (SBOX[(word >> 16) & 0xFF] << 16) |
             (SBOX[(word >> 8) & 0xFF] << 8) |
             SBOX[word & 0xFF];

if you define SBOX as an int[256] array.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/28188#discussion_r2508285414