[clang] [llvm] [Clang][AArch64] Add customisable immediate range checking to NEON (PR #100278)

[via cfe-commits]

================
@@ -1959,9 +2064,12 @@ multiclass VCMLA_ROTS<string type, string lanety, string laneqty> {
 
     let isLaneQ = 1 in  {
       // vcmla{ROT}_laneq
+      // ACLE specifies that the fp16 vcmla_#ROT_laneq variant has an immedaite range of 0 <= lane <= 1.
+      // fp16 is the only variant for which these two differ.
+      // https://developer.arm.com/documentation/ihi0073/latest/ 
+      defvar getlanety = !if(!eq(type, "h"), lanety, laneqty);
       def : SOpInst<"vcmla" # ROT # "_laneq", "...QI", type,  Op<(call "vcmla" # ROT, $p0, $p1,
-              (bitcast $p0, (dup_typed lanety, (call "vget_lane", (bitcast laneqty, $p2), $p3))))>>;
-
+                (bitcast $p0, (dup_typed lanety, (call "vget_lane", (bitcast getlanety, $p2), $p3))))>>;
----------------
Lukacma wrote:

I did take a look into  [ACLE release](https://github.com/ARM-software/acle/releases/tag/r2024Q2) and I think the website you linked might have a mistake there. Looking at advsimd-2023Q2 I see `vcmla[_rot]_laneq_f16` defined like this :
<img width="725" alt="image" src="https://github.com/user-attachments/assets/2a446a5c-a3bf-40fc-9950-6caee95d81a2">

It also makes sense for all of them to have the same imm. range, as the range is only dependant on size and datatype of the 3 operand and not on the sizes and datatypes of the other operands. 

https://github.com/llvm/llvm-project/pull/100278