RFR: 8376164: Optimize AES/ECB/PKCS5Padding implementation using full-message intrinsic stub and parallel RoundKey addition [v5]

[xinyangwu]

On Thu, 26 Feb 2026 14:29:27 GMT, Marc Chevalier <mchevalier at openjdk.org> wrote:

> Another thing is that it was run on a Windows Server 2022, in case the OS can play a role. Run in fastdebug mode. I don't see much more relevant info, sorry.

@marc-chevalier Thanks a lot for the additional details — that was very helpful!

Since this was run on Windows Server 2022, the situation became much clearer. Most of my previous testing had been done on linux-x64, and I had not sufficiently exercised this path on Windows.

I have now reproduced the failure on a Win64 system and identified the root cause. The issue has been fixed in the latest update of the PR.

Could you please try rerunning the test with the updated patch and let me know if the problem still occurs on your side?

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

PR Comment: https://git.openjdk.org/jdk/pull/29385#issuecomment-3972810428