[llvm] [llvm][AArch64][Assembly]: Add FP8FMA assembly and disassembly. (PR #70134)

[via llvm-commits]

================
@@ -10078,3 +10078,93 @@ multiclass sve2p1_tblq<string mnemonic> {
   def _S : sve2p1_permute_vec_elems_q<0b10, 0b110, mnemonic, ZPR32, Z_s>;
   def _D : sve2p1_permute_vec_elems_q<0b11, 0b110, mnemonic, ZPR64, Z_d>;
 }
+
+// FP8 Widening Multiply-Add Long - Indexed Group
+class sve2_fp8_mla_long_by_indexed_elem<bit T, string mnemonic>
+    : I<(outs ZPR16:$Zda),
+      (ins ZPR16:$_Zda, ZPR8:$Zn, ZPR3b8:$Zm, VectorIndexB:$imm4),
+      mnemonic, "\t$Zda, $Zn, $Zm$imm4",
+      "", []>, Sched<[]>{
+  bits<5> Zda;
+  bits<5> Zn;
+  bits<3> Zm;
+  bits<4> imm4;
+  let Inst{31-24} = 0b01100100;
+  let Inst{23} = T;
+  let Inst{22-21} = 0b01;
+  let Inst{20-19} = imm4{3-2};
+  let Inst{18-16} = Zm;
+  let Inst{15-12} = 0b0101;
+  let Inst{11-10} = imm4{1-0};
+  let Inst{9-5} = Zn;
+  let Inst{4-0} = Zda;
+  let Constraints = "$Zda = $_Zda";
+  let DestructiveInstType = DestructiveOther;
+  let ElementSize = ZPR16.ElementSize;
+}
+
+// FP8 Widening Multiply-Add Long Group
+class sve2_fp8_mla_long<bit T, string mnemonic>
----------------
CarolineConcatto wrote:

So, I believe we can have one class for  sve2_fp8_mla_long and sve2_fp8_mla_long_long. 
See section: SVE2 FP8 Widening Multiply-Add
in https://developer.arm.com/documentation/ddi0602/2023-09/Index-by-Encoding/SVE-encodings?lang=en

https://github.com/llvm/llvm-project/pull/70134