[11u] RFR 8206895: aarch64: rework error-prone cmp instuction

[Hohensee, Paul]

I'd say this patch is worth doing on its own, since it catches actual argument size mismatches that would otherwise be silently accepted. Doesn't seem super intrusive either: all it does is add formal argument casts to "u1", and replace some cmp instructions with subs. Effectively syntax-only changes, so low risk.

I don't think we've been backporting all of the aarch64 back end work to 11u, just some of it, and we've (well, I've, so far) done the equivalent x64 work where needed too.

Thanks,
Paul

-----Original Message-----
From: jdk-updates-dev <jdk-updates-dev-retn at openjdk.java.net> on behalf of Andrew Haley <aph at redhat.com>
Date: Tuesday, March 23, 2021 at 9:30 AM
To: "jdk-updates-dev at openjdk.java.net" <jdk-updates-dev at openjdk.java.net>
Subject: RE: [11u] RFR 8206895: aarch64: rework error-prone cmp instuction

On 3/22/21 3:08 PM, Dmitry Chuyko wrote:
> Hello,
>
> Original bug: https://bugs.openjdk.java.net/browse/JDK-8206895
>
> Original patch applies with minor fuzz and requires minor adjustments:
> few rejected changes in macroAssembler_aarch64.cpp and
> stubGenerator_aarch64.cpp were recreated -- all because of already
> ported stuff [1][2][3][4].

I'm tempted to say no to this one, because it's too intrusive. If
that means a ton of enhancements such as JDK-8229351 are more difficult
to backport, is that a bad thing? Otherwise we're basically porting all
of the AArch64 back end from head to 11u.

--
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