[llvm] [RISCV] Initial codegen support for the XRivosVizip extension (PR #131933)

[Min-Yih Hsu via llvm-commits]

================
@@ -67,6 +67,40 @@ defm RI_VUNZIP2A_V : VALU_IV_V<"ri.vunzip2a", 0b001000>;
 defm RI_VUNZIP2B_V : VALU_IV_V<"ri.vunzip2b", 0b011000>;
 }
 
+// These are modeled after the int binop VL nodes
+def ri_vzipeven_vl : SDNode<"RISCVISD::RI_VZIPEVEN_VL", SDT_RISCVIntBinOp_VL>;
+def ri_vzipodd_vl : SDNode<"RISCVISD::RI_VZIPODD_VL", SDT_RISCVIntBinOp_VL>;
+def ri_vzip2a_vl : SDNode<"RISCVISD::RI_VZIP2A_VL", SDT_RISCVIntBinOp_VL>;
+
+multiclass RIVPseudoVALU_VV {
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defm "" : VPseudoBinaryV_VV<m, Commutable=0>;
+  }
+}
+
+let Predicates = [HasVendorXRivosVizip],
+  Constraints = "@earlyclobber $rd, $rd = $passthru" in {
+defm PseudoRI_VZIPEVEN   : RIVPseudoVALU_VV;
+defm PseudoRI_VZIPODD   : RIVPseudoVALU_VV;
+defm PseudoRI_VZIP2A   : RIVPseudoVALU_VV;
+}
+
+multiclass RIVPatBinaryVL_VV<SDPatternOperator vop, string instruction_name,
+                              list<VTypeInfo> vtilist = AllIntegerVectors,
+                              bit isSEWAware = 0> {
----------------
mshockwave wrote:

nit: isSEWAware = false

https://github.com/llvm/llvm-project/pull/131933