[aarch64-port-dev ] [8u] Patches needed in 8u aarch64

[Wang Zhuo(Zhuoren)]

Thanks, if nobody is backporting JDK-8069191 and JDK-8183543, I will do it.

Regards,
Zhuoren


------------------------------------------------------------------
From:Andrew Haley <aph at redhat.com>
Sent At:2021 Mar. 17 (Wed.) 19:33
To:aarch64-port-dev <aarch64-port-dev at openjdk.java.net>
Subject:Re: [aarch64-port-dev ] [8u] Patches needed in 8u aarch64

On 3/17/21 2:34 AM, Wang Zhuo(Zhuoren) wrote:
> Alibaba ported a number of online applications last year, and we found these patches were useful for current 8u aarch64.
> P1: 
> JDK-8069191:moving predicate out of loops may cause array accesses to bypass null check
> Link: https://bugs.openjdk.java.net/browse/JDK-8069191
> This bug caused a lot of applications crashed during our test, easy to reproduce. I think it is a must-have.

yes. That sounds right.

> P2:
> JDK-8155949: Support relaxed semantics in cmpxchg
> https://bugs.openjdk.java.net/browse/JDK-8155949

That patch seems a bit too risky for 8u, for not much gain.

> JDK-8157904: Atomic::cmpxchg for jbyte is missing a fence on initial failure
> https://bugs.openjdk.java.net/browse/JDK-8157904

Are you sure this is useful on AArch64? That code should not be
used on the AArch64 platform.

> JDK-8196064 : AArch64: Merging ld/st into ldp/stp in macro-assembler
> https://bugs.openjdk.java.net/browse/JDK-8196064

It's a nice patch, but maybe too much for 8u? I'm not sure.

> JDK-8183543: Aarch64: C2 compilation often fails with "failed spill-split-recycle sanity check"
> https://bugs.openjdk.java.net/browse/JDK-8183543

Yes.

-- 
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