[Offload] Implement 'Vendor Name' device info for CUDA

[callumfare]

After #146345 the device info implementation requires a value for every query, rather than silently returning an empty string. This broke the test for OL_DEVICE_INFO_VENDOR on CUDA.

Add a value to the CUDA plugin. We can quite safely hard code this one.

[llvmbot]

@llvm/pr-subscribers-offload

Author: Callum Fare (callumfare)

Changes

After #146345 the device info implementation requires a value for every query, rather than silently returning an empty string. This broke the test for OL_DEVICE_INFO_VENDOR on CUDA.

Add a value to the CUDA plugin. We can quite safely hard code this one.

---

Full diff: https://github.com/llvm/llvm-project/pull/147334.diff

1 Files Affected:

• (modified) offload/plugins-nextgen/cuda/src/rtl.cpp (+2)

diff --git a/offload/plugins-nextgen/cuda/src/rtl.cpp b/offload/plugins-nextgen/cuda/src/rtl.cpp
index f1164074f9ea9..536a294f07ce3 100644
--- a/offload/plugins-nextgen/cuda/src/rtl.cpp
+++ b/offload/plugins-nextgen/cuda/src/rtl.cpp
@@ -940,6 +940,8 @@ struct CUDADeviceTy : public GenericDeviceTy {
     if (Res == CUDA_SUCCESS)
       Info.add("Device Name", TmpChar);
 
+    Info.add("Vendor Name", "NVIDIA");
+
     Res = cuDeviceTotalMem(&TmpSt, Device);
     if (Res == CUDA_SUCCESS)
       Info.add("Global Memory Size", TmpSt, "bytes");