[llvm] [OpenMP][Offload] Update the Logic for Configuring Auto Zero-Copy (PR #143638)

[via llvm-commits]

================
@@ -286,16 +286,16 @@ void PluginManager::registerLib(__tgt_bin_desc *Desc) {
   }
   PM->RTLsMtx.unlock();
 
-  bool UseAutoZeroCopy = Plugins.size() > 0;
+  bool UseAutoZeroCopy = false;
 
   auto ExclusiveDevicesAccessor = getExclusiveDevicesAccessor();
-  for (const auto &Device : *ExclusiveDevicesAccessor)
-    UseAutoZeroCopy &= Device->useAutoZeroCopy();
+  // APUs are homogeneous set of GPUs. Check the first device for
+  // configuring Auto Zero-Copy.
+  if (ExclusiveDevicesAccessor->size() > 0) {
+    auto &Device = *(*ExclusiveDevicesAccessor)[0];
+    UseAutoZeroCopy = Device.useAutoZeroCopy();
----------------
carlobertolli wrote:

The plug-in determines whether this is a case of auto zero copy and in case you f the amdgpu plugin it does so by verifying that the node is an APU. We do not abstract the concept of APU, which is vendor specific, to the top level libomptarget.

https://github.com/llvm/llvm-project/pull/143638