[jdk18] RFR: 8278889: AArch64: [vectorapi] VectorMaskLoadStoreTest.testMaskCast() test fail [v2]

[Eric Liu]

On Wed, 22 Dec 2021 02:16:06 GMT, Vladimir Kozlov <kvn at openjdk.org> wrote:

> It is actually [8275316](https://bugs.openjdk.java.net/browse/JDK-8275316) which was closed as duplicate [8275263 ](https://bugs.openjdk.java.net/browse/JDK-8275263). But, as I understand, 8275263 was also closed as duplicate of [BACKOUT] [8275262](https://bugs.openjdk.java.net/browse/JDK-8275262). So there should not be code in JDK 18 which cause it.
> 
> Note, I ran additional testing of the same local repo build but without this **8278889** changes and I don't see `gcbasher` failures.

Thanks for your feedback.

The failure is really bizarre. Is it possible that caused by the specific toolchain ? This patch doesn't change any common code except the test cases, all the changes are SVE related, and I think mac couldn't touch SVE related code at this moment. Moreover, the log is the same as [8275316](https://bugs.openjdk.java.net/browse/JDK-8275316), which means perhaps some other patches will trigger this failure as well before the root cause has been fixed.  In this way, expose this failure and post a new JBS to involve some mac experts maybe not bad. Otherwise, some patches which looks okay but would be blocked unfortunately.

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

PR: https://git.openjdk.java.net/jdk18/pull/49