[llvm] [AArch64] Tighten conditions for expanding GET_ACTIVE_LANE_MASK (PR #208962)

[Kerry McLaughlin via llvm-commits]

================
@@ -494,3 +424,6 @@ declare <16 x i1> @llvm.get.active.lane.mask.v16i1.i8(i8, i8)
 declare <8 x i1> @llvm.get.active.lane.mask.v8i1.i8(i8, i8)
 declare <4 x i1> @llvm.get.active.lane.mask.v4i1.i8(i8, i8)
 declare <2 x i1> @llvm.get.active.lane.mask.v2i1.i8(i8, i8)
+;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
+; CHECK-STREAMING: {{.*}}
+; CHECK-SVE: {{.*}}
----------------
kmclaughlin-arm wrote:

I think the `--check-prefixes` flags can be removed from the RUN lines of this test, since there is no longer a difference in output between the streaming & sve checks?

https://github.com/llvm/llvm-project/pull/208962