[jdk11u-dev] RFR: 8248238: Implementation: JEP 388: Windows AArch64 Support [v12]

[Andrew Dinn]

On Wed, 24 Nov 2021 19:58:55 GMT, Reka Kovacs <rkovacs at openjdk.org> wrote:

>>> [Here](https://gist.github.com/rnkovacs/edfa07b7b86ae720c26e788db230f872)'s a list of the failing hotspot tests.
>>> 
>>> We've been battling issues with both Cygwin and MSVC. We've reported a [bad arm64 codegen bug](https://developercommunity.visualstudio.com/t/arm64-incorrect-code-gen-when-compiler-decides-to/1541140) that causes assert failures in a bunch of tests, and is still present in VS 2022 (closed as fixed incorrectly - we're contacting the engineers directly). We got our results by using a [workaround](https://gist.github.com/rnkovacs/8b67a8513d19b43343185dc5f86fe3fe) that seems to fix the issue.
>> 
>> I don't see any reason why we'd reject your workaround, so please add it to this PR.
>> 
>> I'll approve this _pro tem_  without the Serviceability fixes, but I don't want them to be neglected. Please let us know about your plans to get Serviceability done.
>
>> I don't see any reason why we'd reject your workaround, so please add it to this PR.
>> 
>> I'll approve this _pro tem_ without the Serviceability fixes, but I don't want them to be neglected. Please let us know about your plans to get Serviceability done.
> 
> Thank you - I added it in a new commit.
> 
> I got reassured that fixing these failures is a top priority for us, especially since we see most of them with our JDK 17 builds as well. We can start investigating right away, but we'd like to merge this first to make sure it's not delayed any longer.

@rnkovacs Ok, no problem with that test failure then. Good to merge then.

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

PR: https://git.openjdk.java.net/jdk11u-dev/pull/301