[clang] [llvm] [AArch64][clang][llvm] Add ACLE Armv9.7 matrix multiply-accumulate intrinsics (PR #193017)

[via cfe-commits]

================
@@ -225,7 +225,7 @@ def HasSVE2p2        : Predicate<"Subtarget->hasSVE2p2()">,
                                  AssemblerPredicateWithAll<(all_of FeatureSVE2p2), "sve2p2">;
 def HasSVE_B16MM     : Predicate<"Subtarget->isSVEAvailable() && Subtarget->hasSVE_B16MM()">,
                                  AssemblerPredicateWithAll<(all_of FeatureSVE_B16MM), "sve-b16mm">;
-def HasF16MM         : Predicate<"Subtarget->isSVEAvailable() && Subtarget->hasF16MM()">,
+def HasF16MM         : Predicate<"Subtarget->hasF16MM()">,
----------------
Lukacma wrote:

If I am reading this correctly this change would allow FMMLA_ZZZ_H to be run in streaming mode, which is not correct.

https://github.com/llvm/llvm-project/pull/193017