[llvm] [ISel] Introduce llvm.clmul intrinsic (PR #168731)

[Ramkumar Ramachandra via llvm-commits]

================
@@ -767,6 +767,11 @@ enum NodeType {
   FSHL,
   FSHR,
 
+  /// Carry-less multiplication operations.
+  CLMUL,
+  CLMULR,
----------------
artagnon wrote:

My next patch is for RISC-V, factoring out llvm.riscv.clmul[rh], so I'd prefer to keep this, if you don't mind. The following patch will be to LoopIdiom, getting big-endian/little-endian CRC lowered to RISC-V's clmul[rh] -- without clmul[rh], we wouldn't be able to do that effectively.

https://github.com/llvm/llvm-project/pull/168731