AMDGPU: Mark all unspecified CC functions in tests as amdgpu_kernel

Currently the default C calling convention functions are treated
the same as compute kernels. Make this explicit so the default
calling convention can be changed to a non-kernel.

Converted with perl -pi -e 's/define void/define amdgpu_kernel void/'
on the relevant test directories (and undoing in one place that actually
wanted a non-kernel).

llvm-svn: 298444
diff --git a/llvm/test/CodeGen/AMDGPU/parallelorifcollapse.ll b/llvm/test/CodeGen/AMDGPU/parallelorifcollapse.ll
index 1da1e91b8..91116b0 100644
--- a/llvm/test/CodeGen/AMDGPU/parallelorifcollapse.ll
+++ b/llvm/test/CodeGen/AMDGPU/parallelorifcollapse.ll
@@ -12,7 +12,7 @@
 ; CHECK: OR_INT
 ; CHECK-NEXT: OR_INT
 ; CHECK-NEXT: OR_INT
-define void @_Z9chk1D_512v() #0 {
+define amdgpu_kernel void @_Z9chk1D_512v() #0 {
 entry:
   %a0 = alloca i32, align 4
   %b0 = alloca i32, align 4