[llvm] [Offload][AMDGPU] accept generic target (PR #118919)

[Matt Arsenault via llvm-commits]

================
@@ -3207,13 +3209,17 @@ struct AMDGPUPluginTy final : public GenericPluginTy {
     if (!Processor)
       return false;
 
-    auto TargeTripleAndFeaturesOrError =
-        hsa_utils::getTargetTripleAndFeatures(getKernelAgent(DeviceId));
-    if (!TargeTripleAndFeaturesOrError)
-      return TargeTripleAndFeaturesOrError.takeError();
-    return offloading::amdgpu::isImageCompatibleWithEnv(
-        Processor ? *Processor : "", ElfOrErr->getPlatformFlags(),
-        *TargeTripleAndFeaturesOrError);
+    SmallVector<std::string> Targets;
+    Targets.push_back("");
----------------
arsenm wrote:

```suggestion
    SmallVector<std::string> Targets = {""};
```

https://github.com/llvm/llvm-project/pull/118919