[llvm] Add support for DFP IR type. (PR #69718)

[Andy Kaylor via llvm-commits]

================
@@ -3672,6 +3678,9 @@ Floating-Point Types
    * - ``ppc_fp128``
      - 128-bit floating-point value (two 64-bits)
 
+   * - ``decimal128``
----------------
andykaylor wrote:

[nit] I'd keep this together with the others.

https://github.com/llvm/llvm-project/pull/69718