[RISCV] Fix vmerge.vvm/vmv.v.v getting folded into ops with mismatchi…

…ng EEW (#101152)

As noted in
https://github.com/llvm/llvm-project/pull/100367/files#r1695448771, we
currently fold in vmerge.vvms and vmv.v.vs into their ops even if the
EEW is different which leads to an incorrect transform.

This checks the op's EEW via its simple value type for now since there
doesn't seem to be any existing information about the EEW size of
instructions. We'll probably need to encode this at some point if we
want to be able to access it at the MachineInstr level in #100367

llvm/lib/Target/RISCV/RISCVISelDAGToDAG.cpp

@@ -3733,6 +3733,10 @@ bool RISCVDAGToDAGISel::performCombineVMergeAndVOps(SDNode *N) {
   assert(!Mask || cast<RegisterSDNode>(Mask)->getReg() == RISCV::V0);
   assert(!Glue || Glue.getValueType() == MVT::Glue);
 
+  // If the EEW of True is different from vmerge's SEW, then we can't fold.
+  if (True.getSimpleValueType() != N->getSimpleValueType(0))
+    return false;
+
   // We require that either passthru and false are the same, or that passthru
   // is undefined.
   if (Passthru != False && !isImplicitDef(Passthru))

llvm/test/CodeGen/RISCV/rvv/rvv-peephole-vmerge-vops.ll

@@ -1196,3 +1196,24 @@ define <vscale x 2 x i32> @true_mask_vmerge_implicit_passthru(<vscale x 2 x i32>
   )
   ret <vscale x 2 x i32> %b
 }
+
+
+define <vscale x 2 x i32> @unfoldable_mismatched_sew(<vscale x 2 x i32> %passthru, <vscale x 1 x i64> %x, <vscale x 1 x i64> %y, <vscale x 2 x i1> %mask, i64 %avl) {
+; CHECK-LABEL: unfoldable_mismatched_sew:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m1, ta, ma
+; CHECK-NEXT:    vadd.vv v9, v9, v10
+; CHECK-NEXT:    vsetvli zero, a0, e32, m1, tu, ma
+; CHECK-NEXT:    vmv.v.v v8, v9
+; CHECK-NEXT:    ret
+  %a = call <vscale x 1 x i64> @llvm.riscv.vadd.nxv1i64.nxv1i64(<vscale x 1 x i64> poison, <vscale x 1 x i64> %x, <vscale x 1 x i64> %y, i64 %avl)
+  %a.bitcast = bitcast <vscale x 1 x i64> %a to <vscale x 2 x i32>
+  %b = call <vscale x 2 x i32> @llvm.riscv.vmerge.nxv2i32.nxv2i32(
+    <vscale x 2 x i32> %passthru,
+    <vscale x 2 x i32> %passthru,
+    <vscale x 2 x i32> %a.bitcast,
+    <vscale x 2 x i1> splat (i1 true),
+    i64 %avl
+  )
+  ret <vscale x 2 x i32> %b
+}

llvm/test/CodeGen/RISCV/rvv/vmv.v.v-peephole.ll

@@ -180,3 +180,17 @@ define <vscale x 2 x i32> @unfoldable_vredsum(<vscale x 2 x i32> %passthru, <vsc
   %b = call <vscale x 2 x i32> @llvm.riscv.vmv.v.v.nxv2i32(<vscale x 2 x i32> %passthru, <vscale x 2 x i32> %a, iXLen 1)
   ret <vscale x 2 x i32> %b
 }
+
+define <vscale x 2 x i32> @unfoldable_mismatched_sew(<vscale x 2 x i32> %passthru, <vscale x 1 x i64> %x, <vscale x 1 x i64> %y, iXLen %avl) {
+; CHECK-LABEL: unfoldable_mismatched_sew:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m1, ta, ma
+; CHECK-NEXT:    vadd.vv v9, v9, v10
+; CHECK-NEXT:    vsetvli zero, a0, e32, m1, tu, ma
+; CHECK-NEXT:    vmv.v.v v8, v9
+; CHECK-NEXT:    ret
+  %a = call <vscale x 1 x i64> @llvm.riscv.vadd.nxv1i64.nxv1i64(<vscale x 1 x i64> poison, <vscale x 1 x i64> %x, <vscale x 1 x i64> %y, iXLen %avl)
+  %a.bitcast = bitcast <vscale x 1 x i64> %a to <vscale x 2 x i32>
+  %b = call <vscale x 2 x i32> @llvm.riscv.vmv.v.v.nxv2i32(<vscale x 2 x i32> %passthru, <vscale x 2 x i32> %a.bitcast, iXLen %avl)
+  ret <vscale x 2 x i32> %b
+}