[llvm] [VPlan] Emit VPVectorEndPointerRecipe for reverse interleave pointer adjustment (PR #144864)

[Mel Chen via llvm-commits]

================
@@ -334,22 +334,24 @@ define void @vector_reverse_i64(ptr nocapture noundef writeonly %A, ptr nocaptur
 ; CHECK-NEXT:    %22 = zext i32 %21 to i64
 ; CHECK-NEXT:    %23 = getelementptr inbounds i32, ptr %B, i64 %22
 ; CHECK-NEXT:    %24 = mul i64 0, %18
-; CHECK-NEXT:    %25 = sub i64 1, %18
-; CHECK-NEXT:    %26 = getelementptr inbounds i32, ptr %23, i64 %24
-; CHECK-NEXT:    %27 = getelementptr inbounds i32, ptr %26, i64 %25
-; CHECK-NEXT:    %wide.load = load <vscale x 4 x i32>, ptr %27, align 4
+; CHECK-NEXT:    %25 = sub i64 %18, 1
----------------
Mel-Chen wrote:

Yeah, I often spend quite a bit of time manually updating this test case too.
But I'm not sure what this test was originally intended to verify — is it the generated IR, the shape of the VPlan, or something else that needs to be checked? cc. @LiqinWeng 

https://github.com/llvm/llvm-project/pull/144864