Add amdgpu to gpu-kernel test

Check that the gpu-kernel calling convention translates to
`amdgpu_kernel` when compiling for the amdgpu target.

tests/codegen-llvm/gpu-kernel-abi.rs

@@ -1,7 +1,9 @@
 // Checks that the gpu-kernel calling convention correctly translates to LLVM calling conventions.
 
 //@ add-core-stubs
-//@ revisions: nvptx
+//@ revisions: amdgpu nvptx
+//@ [amdgpu] compile-flags: --crate-type=rlib --target=amdgcn-amd-amdhsa -Ctarget-cpu=gfx900
+//@ [amdgpu] needs-llvm-components: amdgpu
 //@ [nvptx] compile-flags: --crate-type=rlib --target=nvptx64-nvidia-cuda
 //@ [nvptx] needs-llvm-components: nvptx
 #![feature(no_core, lang_items, abi_gpu_kernel)]
@@ -10,6 +12,7 @@
 extern crate minicore;
 use minicore::*;
 
+// amdgpu: define amdgpu_kernel void @fun(i32
 // nvptx: define ptx_kernel void @fun(i32
 #[no_mangle]
 pub extern "gpu-kernel" fn fun(_: i32) {}