[llvm] [NVPTX] Support address offsets added with disjoint or (PR #122042)

[Kevin McAfee via llvm-commits]

================
@@ -0,0 +1,25 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
+; RUN: llc < %s -mtriple=nvptx64 -mcpu=sm_35 -verify-machineinstrs | FileCheck %s
+; RUN: %if ptxas %{ llc < %s -mtriple=nvptx64 -mcpu=sm_35 | %ptxas-verify %}
+target triple = "nvptx64-nvidia-cuda"
+
+ at a = external global ptr align 16
+
+define i32  @test_v2i8(i16 %a) {
----------------
kalxr wrote:

v2i8?

https://github.com/llvm/llvm-project/pull/122042