[clang] [llvm] [Offload][CUDA] Allow CUDA kernels to use LLVM/Offload (PR #94549)

[via cfe-commits]

github-actions[bot] wrote:

<!--LLVM CODE FORMAT COMMENT: {clang-format}-->


:warning: C/C++ code formatter, clang-format found issues in your code. :warning:

<details>
<summary>
You can test this locally with the following command:
</summary>

``````````bash
git-clang-format --diff bd815a54899d7fa83f6fb49d86c417e1e2e4d2ef 3d5c61a78e91ecb379a2bfac71988eaf8e5cd9cd -- clang/lib/Headers/openmp_wrappers/__llvm_offload.h clang/lib/Headers/openmp_wrappers/__llvm_offload_device.h clang/lib/Headers/openmp_wrappers/__llvm_offload_host.h offload/src/KernelLanguage/API.cpp clang/lib/CodeGen/CGCUDANV.cpp clang/lib/Driver/Driver.cpp clang/lib/Driver/ToolChains/Clang.cpp clang/lib/Driver/ToolChains/CommonArgs.cpp clang/lib/Headers/openmp_wrappers/__clang_openmp_device_functions.h clang/lib/Sema/SemaCUDA.cpp offload/include/Shared/APITypes.h offload/include/omptarget.h offload/plugins-nextgen/amdgpu/src/rtl.cpp offload/plugins-nextgen/common/src/PluginInterface.cpp
``````````

</details>

<details>
<summary>
View the diff from clang-format here.
</summary>

``````````diff
diff --git a/offload/src/KernelLanguage/API.cpp b/offload/src/KernelLanguage/API.cpp
index 07517ad1fc..032f8dc38c 100644
--- a/offload/src/KernelLanguage/API.cpp
+++ b/offload/src/KernelLanguage/API.cpp
@@ -8,7 +8,6 @@
 //
 //===----------------------------------------------------------------------===//
 
-
 #include "Shared/APITypes.h"
 
 #include <cstdio>
@@ -45,7 +44,7 @@ unsigned __llvmPushCallConfiguration(dim3 __grid_size, dim3 __block_size,
 
 unsigned __llvmPopCallConfiguration(dim3 *__grid_size, dim3 *__block_size,
                                     size_t *__shared_memory, void *__stream) {
-   __omp_kernel_t &__kernel = __current_kernel;
+  __omp_kernel_t &__kernel = __current_kernel;
   *__grid_size = __kernel.__grid_size;
   *__block_size = __kernel.__block_size;
   *__shared_memory = __kernel.__shared_memory;
@@ -69,8 +68,7 @@ unsigned llvmLaunchKernel(const void *func, dim3 gridDim, dim3 blockDim,
   Args.ThreadLimit[2] = blockDim.z;
   Args.ArgPtrs = args;
   Args.Flags.IsCUDA = true;
-  int rv = __tgt_target_kernel(nullptr, 0, gridDim.x,
-                               blockDim.x, func, &Args);
+  int rv = __tgt_target_kernel(nullptr, 0, gridDim.x, blockDim.x, func, &Args);
   return rv;
 }
 }

``````````

</details>


https://github.com/llvm/llvm-project/pull/94549