[llvm] [SelectionDAG] WidenVecOp_INSERT_SUBVECTOR - Replace `INSERT_SUBVECTOR` with series of `INSERT_VECTOR_ELT` (PR #124420)

[Simon Pilgrim via llvm-commits]

================
@@ -0,0 +1,47 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
+; RUN: llc < %s -mtriple=x86_64-- | FileCheck %s
+
+define <4 x i32> @insert_i32_v2_in_v4_at_0(<4 x i32> %a, <2 x i32> %b) {
+; CHECK-LABEL: insert_i32_v2_in_v4_at_0:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; CHECK-NEXT:    retq
+  %result = tail call <4 x i32> @llvm.vector.insert.v4i32.v2i32(<4 x i32> %a, <2 x i32> %b, i64 0)
+  ret <4 x i32> %result
+}
+
+define <4 x i32> @insert_i32_v2_in_v4_at_2(<4 x i32> %a, <2 x i32> %b) {
+; CHECK-LABEL: insert_i32_v2_in_v4_at_2:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; CHECK-NEXT:    retq
+  %result = tail call <4 x i32> @llvm.vector.insert.v4i32.v2i32(<4 x i32> %a, <2 x i32> %b, i64 2)
+  ret <4 x i32> %result
+}
+
+define <4 x float> @insert_f32_v2_in_v4_at_0(<4 x float> %a, <2 x float> %b) {
+; CHECK-LABEL: insert_f32_v2_in_v4_at_0:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; CHECK-NEXT:    retq
+  %result = tail call <4 x float> @llvm.vector.insert.v4float.v2float(<4 x float> %a, <2 x float> %b, i64 0)
+  ret <4 x float> %result
+}
+
+define <8 x i32> @insert_i32_v2_in_v8_at_0(<8 x i32> %a, <2 x i32> %b) {
+; CHECK-LABEL: insert_i32_v2_in_v8_at_0:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    movsd {{.*#+}} xmm0 = xmm2[0],xmm0[1]
+; CHECK-NEXT:    retq
+  %result = tail call <8 x i32> @llvm.vector.insert.v8i32.v2i32(<8 x i32> %a, <2 x i32> %b, i64 0)
+  ret <8 x i32> %result
+}
+
+define <8 x i32> @insert_i32_v2_in_v8_at_6(<8 x i32> %a, <2 x i32> %b) {
+; CHECK-LABEL: insert_i32_v2_in_v8_at_6:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    movlhps {{.*#+}} xmm1 = xmm1[0],xmm2[0]
+; CHECK-NEXT:    retq
+  %result = tail call <8 x i32> @llvm.vector.insert.v8i32.v2i32(<8 x i32> %a, <2 x i32> %b, i64 6)
+  ret <8 x i32> %result
+}
----------------
RKSimon wrote:

Ideally we should fix the AVX case as well - it'd be OK to keep the tests for now and add -mattr=+avx test coverage along with the fix in a future PR

https://github.com/llvm/llvm-project/pull/124420