[compiler-rt] [libunwind] [AArch64] Fix nofp regressions in compiler-rt and libunwind (PR #111235)

[Keith Packard via cfe-commits]

================
@@ -1,5 +1,20 @@
 #include <stddef.h>
 
+#if __ARM_FP == 0
+// WARNING: When building the scalar versions of these functions you need to
+// use the compiler flag "-mllvm -disable-loop-idiom-all" to prevent clang
----------------
keith-packard wrote:

I just checked -- the CMakeLists.txt file already uses `-fno-builtins`:
```    
set_source_files_properties(aarch64/sme-libc-routines.c PROPERTIES COMPILE_FLAGS "-fno-builtin")
```
I've removed this mis-leading comment; we need `-fno-builtin` for every function in this file, and the memchr version doesn't talk about it.

https://github.com/llvm/llvm-project/pull/111235