[clang] [Clang][GPU] Make NVPTX check more permissive in unit test (PR #136301)

[via cfe-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-clang

Author: Rahul Joshi (jurahul)

<details>
<summary>Changes</summary>

- Seems based on whether NVPTX backend is enabled or not, this call can have the range() attribute or not. So make this check more permissive.

---
Full diff: https://github.com/llvm/llvm-project/pull/136301.diff


1 Files Affected:

- (modified) clang/test/Headers/gpuintrin_lang.c (+1-1) 


``````````diff
diff --git a/clang/test/Headers/gpuintrin_lang.c b/clang/test/Headers/gpuintrin_lang.c
index b804d46071507..653f87aea2ce3 100644
--- a/clang/test/Headers/gpuintrin_lang.c
+++ b/clang/test/Headers/gpuintrin_lang.c
@@ -36,7 +36,7 @@ __device__ int foo() { return __gpu_thread_id_x(); }
 // CUDA-LABEL: define dso_local i32 @foo(
 // CUDA-SAME: ) #[[ATTR0:[0-9]+]] {
 // CUDA-NEXT:  [[ENTRY:.*:]]
-// CUDA-NEXT:    [[TMP0:%.*]] = call i32 @llvm.nvvm.read.ptx.sreg.tid.x()
+// CUDA-NEXT:    [[TMP0:%.*]] = call {{.*}}i32 @llvm.nvvm.read.ptx.sreg.tid.x()
 // CUDA-NEXT:    ret i32 [[TMP0]]
 //
 // HIP-LABEL: define dso_local i32 @foo(

``````````

</details>


https://github.com/llvm/llvm-project/pull/136301