[llvm] [llvm] Preserve existing calling conv when auto-upgrading nvvm.annota… (PR #125568)

[Shilei Tian via llvm-commits]

================
@@ -0,0 +1,14 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
+; RUN: opt < %s -mtriple=amdgcn-amd-amdhsa -O0 -S | FileCheck %s
+
+define amdgpu_kernel void @foo(i32 %a, i32 %b) {
+; CHECK-LABEL: define amdgpu_kernel void @foo(
+; CHECK-SAME: i32 [[A:%.*]], i32 [[B:%.*]]) {
+; CHECK-NEXT:    ret void
+;
+  ret void
+}
+
+!nvvm.annotations = !{!0}
----------------
shiltian wrote:

Why do we have `nvvm` annotation along with `amdgpu_kernel` at the first place?

https://github.com/llvm/llvm-project/pull/125568