On Tue, 10 Nov 2026 21:36:55 GMT, Mohamed Issa <[email protected]> wrote:
>> The primary reason for this set of changes is to sanity check the >> **doubleKeccak** and **quadKeccak** methods. This includes their >> intrinsified and pure Java variants. The new JTREG source files listed below >> essentially validate return values from invoked private methods using the >> Java Reflection API. >> >> 1. `jtreg:test/hotspot/jtreg/compiler/intrinsics/sha/TestInvoke.java` >> 2. >> `jtreg:test/hotspot/jtreg/compiler/intrinsics/sha/sanity/InvokeSanityTestBase.java` >> 3. >> `jtreg:test/hotspot/jtreg/compiler/intrinsics/sha/sanity/TestSHA3ParallelIntrinsics.java` >> >> Additionally, the multi block intrinsic tests now check the correct >> **implCompressMultiBlock0** method. All modifications and tests used >> [OpenJDK v28-b22](https://github.com/openjdk/jdk/releases/tag/jdk-28%2B22) >> as the baseline build. >> >> --------- >> - [x] I confirm that I make this contribution in accordance with the >> [OpenJDK Interim AI Policy](https://openjdk.org/legal/ai). > > Mohamed Issa has updated the pull request incrementally with three additional > commits since the last revision: > > - Remove check for _quad_keccak stub as that change should be done in > separate PR. > - Use AND predicates in multi-block intrinsic tests in case a VM setting > makes stubs silently unavailable. > - Remove and block tiered compilation from SHA3 parallel tests as that > conflicts with compile threshold setting. > I'm still seeing this failure on macosx-aarch64-debug: > > ``` > class = sun.security.provider.SHA3Parallel > method = quadKeccak > TestInvoke Error: > expected = 0 > actual = 1 > ----------System.err:(17/1127)---------- > java.lang.Exception: Exception: java.lang.Exception: TestInvoke Error > at compiler.intrinsics.sha.TestInvoke.testInvoke(TestInvoke.java:98) ... I believe that this is expected behavior before the fix for JDK-8397115. ------------- PR Comment: https://git.openjdk.org/jdk/pull/33709#issuecomment-6816931070
