[clang] [llvm] [mlir] [AArch64][SME] Improve codegen for aarch64.sme.cnts* when not in streaming mode (PR #154761)

[Paul Walker via llvm-commits]

================
@@ -127,11 +127,21 @@ def : Pat<(AArch64_requires_za_save), (RequiresZASavePseudo)>;
 def SDT_AArch64RDSVL  : SDTypeProfile<1, 1, [SDTCisInt<0>, SDTCisInt<1>]>;
 def AArch64rdsvl : SDNode<"AArch64ISD::RDSVL", SDT_AArch64RDSVL>;
 
+def sme_cntsb_mul_imm : ComplexPattern<i64, 1, "SelectRDVLImm<1, 31, 8>">;
+def sme_cntsh_mul_imm : ComplexPattern<i64, 1, "SelectRDVLImm<1, 31, 4>">;
+def sme_cntsw_mul_imm : ComplexPattern<i64, 1, "SelectRDVLImm<1, 31, 2>">;
+def sme_cntsd_mul_imm : ComplexPattern<i64, 1, "SelectRDVLImm<1, 31, 1>">;
----------------
paulwalker-arm wrote:

These are no longer required?

https://github.com/llvm/llvm-project/pull/154761