(8u) RFR: 8131062: aarch64: add support for GHASH acceleration

[Andrew Haley]

On 8/16/21 6:59 PM, Liu, Xin wrote:

> I'd like to request a review of 8131062 for jdk8u.  This patch can
> accelerate AES/GCM 4~5 times on armv8 by leveraging NEON isntructions.
> It can't apply to jdkj8u/hotspot cleanly, but it's trivial to integrate
> it. I just adjust code locations.

I'm thinking of approving this, even though 8u is in long-term maintenance
mode so such enhancements wouldn't normally be appropriate.

Here's my thinking: not only is AES horribly slow in software, it is also
vulnerable to timing attacks. In addition, encryption is a significant
consumer of time and power, and bottlenecks some applications. For that
reason, this patch should go in now, even though it wasn't back ported at
the time because 8u was already a maintenance release.

However, this is an old version of the GHASH intrinsic. You really want
8134869: AARCH64: GHASH intrinsic is not optimal.
https://bugs.openjdk.java.net/browse/JDK-8134869

It'll need a jdk8u-fix-request tag.

-- 
Andrew Haley  (he/him)
Java Platform Lead Engineer
Red Hat UK Ltd. <https://www.redhat.com>
https://keybase.io/andrewhaley
EAC8 43EB D3EF DB98 CC77 2FAD A5CD 6035 332F A671