[clang] [CIR][AArch64] Handle constrained Neon FMA and sqrt (PR #218307)

[Yair Ben Avraham via cfe-commits]

yairbenavraham wrote:

@banach-space Would it make sense to move the remaining three strict FMA tests from `neon-intrinsics-constrained.c` into `neon/fused-multiply-constrained.c` and remove their old blocks? Their unconstrained behavior is already tested in `neon/fused-multiply.c`, so I wouldn’t repeat those checks.

https://github.com/llvm/llvm-project/pull/218307