RFR: 8373059: Test sun/security/provider/acvp/ML_DSA_Intrinsic_Test.java should pass on Aarch64

[Weijun Wang]

On Tue, 9 Dec 2025 14:41:02 GMT, Ferenc Rakoczi <duke at openjdk.org> wrote:

> …hould pass on Aarch64
> 
> The test used to fail because it had checked a stronger equivalence of the results of the Java method and its intrinsified version.
> Other then fixing that, I did some formatting and corrected a comment.

test/jdk/sun/security/provider/pqc/ML_DSA_Intrinsic_Test.java line 49:

> 47: // To run manually:
> 48: // java --add-opens java.base/sun.security.provider=ALL-UNNAMED
> 49: // --add-exports java.base/sun.security.provider=ALL-UNNAMED

Please indent one space to align with lines below.

test/jdk/sun/security/provider/pqc/ML_DSA_Intrinsic_Test.java line 51:

> 49: // --add-exports java.base/sun.security.provider=ALL-UNNAMED
> 50: //  -XX:+UnlockDiagnosticVMOptions -XX:+UseDilithiumIntrinsics
> 51: //  test/jdk/sun/security/provider/acvp/ML_DSA_Intrinsic_Test.java

You've modified the test path above.

test/jdk/sun/security/provider/pqc/ML_DSA_Intrinsic_Test.java line 140:

> 138:         for (int j = 0; j < ML_DSA_N; j++) {
> 139:             coeffs1[j] = rnd.nextInt(2 * ML_DSA_Q) - ML_DSA_Q;
> 140:             coeffs2[j] = rnd.nextInt(2 * ML_DSA_Q) - ML_DSA_Q;

Why are both so small? Maybe only one is enough?

test/jdk/sun/security/provider/pqc/ML_DSA_Intrinsic_Test.java line 147:

> 145: 
> 146:         if (!Arrays.equals(prod1, prod2)) {
> 147:             boolean modQequal = true;

You can copy the "the result is greater than -MONT_Q and less than MONT_Q" comment here.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/28722#discussion_r2614055156
PR Review Comment: https://git.openjdk.org/jdk/pull/28722#discussion_r2614055825
PR Review Comment: https://git.openjdk.org/jdk/pull/28722#discussion_r2614063686
PR Review Comment: https://git.openjdk.org/jdk/pull/28722#discussion_r2614058951