[llvm] [X86][DAGCombiner][SelectionDAG] - Fold Zext Build Vector to Bitcast of widen Build Vector (PR #135010)

[Phoebe Wang via llvm-commits]

================
@@ -799,7 +799,9 @@ define i32 @PR46586(ptr %p, <4 x i32> %v) {
 ;
 ; SSE41-LABEL: PR46586:
 ; SSE41:       # %bb.0:
-; SSE41-NEXT:    movzbl 3(%rdi), %eax
+; SSE41-NEXT:    pxor %xmm1, %xmm1
+; SSE41-NEXT:    pinsrb $12, 3(%rdi), %xmm1
+; SSE41-NEXT:    pextrd $3, %xmm1, %eax
----------------
phoebewang wrote:

Seems also regress on X86?

https://github.com/llvm/llvm-project/pull/135010