[llvm] [Offload] Support loading CUDA fat binaries (PR #156955)

[Jonas Greifenhain via llvm-commits]

================
@@ -556,7 +657,19 @@ struct CUDADeviceTy : public GenericDeviceTy {
 
     // Allocate and initialize the image object.
     CUDADeviceImageTy *CUDAImage = Plugin.allocate<CUDADeviceImageTy>();
-    new (CUDAImage) CUDADeviceImageTy(ImageId, *this, TgtImage);
+
+    uint32_t Magic = *reinterpret_cast<const uint32_t *>(TgtImage->ImageStart);
+    if (Magic == 0x466243b1 || Magic == 0xba55ed50) {
+      // It's a fatbin or a wrapped fatbin
----------------
cadivus wrote:

The way I got it, they wrap it at runtime and have it unwrapped if you store it as a file. 
I think dealing with both here is the easier.

https://github.com/llvm/llvm-project/pull/156955