[llvm] Add --offoading option to llvm-readobj (PR #143342)

[James Henderson via llvm-commits]

================
@@ -0,0 +1,29 @@
+## Test that --offloading with a fatbin works correctly.
+# REQUIRES: target={{x86_64-.*-linux.*}}
+# REQUIRES: amdgpu-registered-target
+# UNSUPPORTED: system-windows
----------------
jh7370 wrote:

I've skimmed the relevant yaml2obj code and I can't see anything immediately that obviously caps the size (especially for Windows only). I actually decided to check out your branch on my Windows machine and look into it.

It turns out that this test passes for me, without any of the `REQUIRES/UNSUPPORTED` directives in this file: I don't have AMDGPU configured and it's building and running fine on Windows.

https://github.com/llvm/llvm-project/pull/143342