[llvm] [AArch64][machine-scheduler][Neoverse-N2] fdiv is blocking (PR #119206)

[Cullen Rhodes via llvm-commits]

================
@@ -795,22 +831,26 @@ def : SchedAlias<WriteF,     N2Write_2c_1V>;
 // FP compare
 def : SchedAlias<WriteFCmp,  N2Write_2c_1V0>;
 
+// FP divide and square root operations are performed using an iterative
+// algorithm and block subsequent similar operations to the same pipeline
+// until complete.
----------------
c-rhodes wrote:

Cite where in optimisation guide this comes from?
```suggestion
// FP divide and square root operations are performed using an iterative
// algorithm and block subsequent similar operations to the same pipeline
// until complete (Arm Neoverse N2 Software Optimization Guide, 3.14).
```

https://github.com/llvm/llvm-project/pull/119206