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/fptrunc.ll b/llvm/test/CodeGen/AMDGPU/fptrunc.ll
index 0c7b674..d9c5b7e 100644
--- a/llvm/test/CodeGen/AMDGPU/fptrunc.ll
+++ b/llvm/test/CodeGen/AMDGPU/fptrunc.ll
@@ -4,7 +4,7 @@
 
 ; FUNC-LABEL: {{^}}fptrunc_f64_to_f32:
 ; GCN: v_cvt_f32_f64_e32 {{v[0-9]+}}, {{s\[[0-9]+:[0-9]+\]}}
-define void @fptrunc_f64_to_f32(float addrspace(1)* %out, double %in) {
+define amdgpu_kernel void @fptrunc_f64_to_f32(float addrspace(1)* %out, double %in) {
   %result = fptrunc double %in to float
   store float %result, float addrspace(1)* %out
   ret void
@@ -14,7 +14,7 @@
 ; GCN-NOT: v_cvt
 ; GCN-UNSAFE: v_cvt_f32_f64_e32 [[F32:v[0-9]+]]
 ; GCN-UNSAFE: v_cvt_f16_f32_e32 v{{[0-9]+}}, [[F32]]
-define void @fptrunc_f64_to_f16(i16 addrspace(1)* %out, double %in) {
+define amdgpu_kernel void @fptrunc_f64_to_f16(i16 addrspace(1)* %out, double %in) {
   %result = fptrunc double %in to half
   %result_i16 = bitcast half %result to i16
   store i16 %result_i16, i16 addrspace(1)* %out
@@ -24,7 +24,7 @@
 ; FUNC-LABEL: {{^}}fptrunc_v2f64_to_v2f32:
 ; GCN: v_cvt_f32_f64_e32
 ; GCN: v_cvt_f32_f64_e32
-define void @fptrunc_v2f64_to_v2f32(<2 x float> addrspace(1)* %out, <2 x double> %in) {
+define amdgpu_kernel void @fptrunc_v2f64_to_v2f32(<2 x float> addrspace(1)* %out, <2 x double> %in) {
   %result = fptrunc <2 x double> %in to <2 x float>
   store <2 x float> %result, <2 x float> addrspace(1)* %out
   ret void
@@ -35,7 +35,7 @@
 ; GCN: v_cvt_f32_f64_e32
 ; GCN: v_cvt_f32_f64_e32
 ; GCN: v_cvt_f32_f64_e32
-define void @fptrunc_v4f64_to_v4f32(<4 x float> addrspace(1)* %out, <4 x double> %in) {
+define amdgpu_kernel void @fptrunc_v4f64_to_v4f32(<4 x float> addrspace(1)* %out, <4 x double> %in) {
   %result = fptrunc <4 x double> %in to <4 x float>
   store <4 x float> %result, <4 x float> addrspace(1)* %out
   ret void
@@ -50,7 +50,7 @@
 ; GCN: v_cvt_f32_f64_e32
 ; GCN: v_cvt_f32_f64_e32
 ; GCN: v_cvt_f32_f64_e32
-define void @fptrunc_v8f64_to_v8f32(<8 x float> addrspace(1)* %out, <8 x double> %in) {
+define amdgpu_kernel void @fptrunc_v8f64_to_v8f32(<8 x float> addrspace(1)* %out, <8 x double> %in) {
   %result = fptrunc <8 x double> %in to <8 x float>
   store <8 x float> %result, <8 x float> addrspace(1)* %out
   ret void