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

1 files changed, 9 insertions, 9 deletions

diff --git a/llvm/test/Analysis/CostModel/AMDGPU/fmul.ll b/llvm/test/Analysis/CostModel/AMDGPU/fmul.ll
index 6303bb7988c..915c35a23b3 100644
--- a/llvm/test/Analysis/CostModel/AMDGPU/fmul.ll
+++ b/llvm/test/Analysis/CostModel/AMDGPU/fmul.ll
@@ -3,7 +3,7 @@
 
 ; ALL: 'fmul_f32'
 ; ALL: estimated cost of 1 for {{.*}} fmul float
-define void @fmul_f32(float addrspace(1)* %out, float addrspace(1)* %vaddr, float %b) #0 {
+define amdgpu_kernel void @fmul_f32(float addrspace(1)* %out, float addrspace(1)* %vaddr, float %b) #0 {
   %vec = load float, float addrspace(1)* %vaddr
   %add = fmul float %vec, %b
   store float %add, float addrspace(1)* %out
@@ -12,7 +12,7 @@ define void @fmul_f32(float addrspace(1)* %out, float addrspace(1)* %vaddr, floa
 
 ; ALL: 'fmul_v2f32'
 ; ALL: estimated cost of 2 for {{.*}} fmul <2 x float>
-define void @fmul_v2f32(<2 x float> addrspace(1)* %out, <2 x float> addrspace(1)* %vaddr, <2 x float> %b) #0 {
+define amdgpu_kernel void @fmul_v2f32(<2 x float> addrspace(1)* %out, <2 x float> addrspace(1)* %vaddr, <2 x float> %b) #0 {
   %vec = load <2 x float>, <2 x float> addrspace(1)* %vaddr
   %add = fmul <2 x float> %vec, %b
   store <2 x float> %add, <2 x float> addrspace(1)* %out
@@ -21,7 +21,7 @@ define void @fmul_v2f32(<2 x float> addrspace(1)* %out, <2 x float> addrspace(1)
 
 ; ALL: 'fmul_v3f32'
 ; ALL: estimated cost of 3 for {{.*}} fmul <3 x float>
-define void @fmul_v3f32(<3 x float> addrspace(1)* %out, <3 x float> addrspace(1)* %vaddr, <3 x float> %b) #0 {
+define amdgpu_kernel void @fmul_v3f32(<3 x float> addrspace(1)* %out, <3 x float> addrspace(1)* %vaddr, <3 x float> %b) #0 {
   %vec = load <3 x float>, <3 x float> addrspace(1)* %vaddr
   %add = fmul <3 x float> %vec, %b
   store <3 x float> %add, <3 x float> addrspace(1)* %out
@@ -31,7 +31,7 @@ define void @fmul_v3f32(<3 x float> addrspace(1)* %out, <3 x float> addrspace(1)
 ; ALL: 'fmul_f64'
 ; FASTF64: estimated cost of 2 for {{.*}} fmul double
 ; SLOWF64: estimated cost of 3 for {{.*}} fmul double
-define void @fmul_f64(double addrspace(1)* %out, double addrspace(1)* %vaddr, double %b) #0 {
+define amdgpu_kernel void @fmul_f64(double addrspace(1)* %out, double addrspace(1)* %vaddr, double %b) #0 {
   %vec = load double, double addrspace(1)* %vaddr
   %add = fmul double %vec, %b
   store double %add, double addrspace(1)* %out
@@ -41,7 +41,7 @@ define void @fmul_f64(double addrspace(1)* %out, double addrspace(1)* %vaddr, do
 ; ALL: 'fmul_v2f64'
 ; FASTF64: estimated cost of 4 for {{.*}} fmul <2 x double>
 ; SLOWF64: estimated cost of 6 for {{.*}} fmul <2 x double>
-define void @fmul_v2f64(<2 x double> addrspace(1)* %out, <2 x double> addrspace(1)* %vaddr, <2 x double> %b) #0 {
+define amdgpu_kernel void @fmul_v2f64(<2 x double> addrspace(1)* %out, <2 x double> addrspace(1)* %vaddr, <2 x double> %b) #0 {
   %vec = load <2 x double>, <2 x double> addrspace(1)* %vaddr
   %add = fmul <2 x double> %vec, %b
   store <2 x double> %add, <2 x double> addrspace(1)* %out
@@ -51,7 +51,7 @@ define void @fmul_v2f64(<2 x double> addrspace(1)* %out, <2 x double> addrspace(
 ; ALL: 'fmul_v3f64'
 ; FASTF64: estimated cost of 6 for {{.*}} fmul <3 x double>
 ; SLOWF64: estimated cost of 9 for {{.*}} fmul <3 x double>
-define void @fmul_v3f64(<3 x double> addrspace(1)* %out, <3 x double> addrspace(1)* %vaddr, <3 x double> %b) #0 {
+define amdgpu_kernel void @fmul_v3f64(<3 x double> addrspace(1)* %out, <3 x double> addrspace(1)* %vaddr, <3 x double> %b) #0 {
   %vec = load <3 x double>, <3 x double> addrspace(1)* %vaddr
   %add = fmul <3 x double> %vec, %b
   store <3 x double> %add, <3 x double> addrspace(1)* %out
@@ -60,7 +60,7 @@ define void @fmul_v3f64(<3 x double> addrspace(1)* %out, <3 x double> addrspace(
 
 ; ALL 'fmul_f16'
 ; ALL estimated cost of 1 for {{.*}} fmul half
-define void @fmul_f16(half addrspace(1)* %out, half addrspace(1)* %vaddr, half %b) #0 {
+define amdgpu_kernel void @fmul_f16(half addrspace(1)* %out, half addrspace(1)* %vaddr, half %b) #0 {
   %vec = load half, half addrspace(1)* %vaddr
   %add = fmul half %vec, %b
   store half %add, half addrspace(1)* %out
@@ -69,7 +69,7 @@ define void @fmul_f16(half addrspace(1)* %out, half addrspace(1)* %vaddr, half %
 
 ; ALL 'fmul_v2f16'
 ; ALL estimated cost of 2 for {{.*}} fmul <2 x half>
-define void @fmul_v2f16(<2 x half> addrspace(1)* %out, <2 x half> addrspace(1)* %vaddr, <2 x half> %b) #0 {
+define amdgpu_kernel void @fmul_v2f16(<2 x half> addrspace(1)* %out, <2 x half> addrspace(1)* %vaddr, <2 x half> %b) #0 {
   %vec = load <2 x half>, <2 x half> addrspace(1)* %vaddr
   %add = fmul <2 x half> %vec, %b
   store <2 x half> %add, <2 x half> addrspace(1)* %out
@@ -78,7 +78,7 @@ define void @fmul_v2f16(<2 x half> addrspace(1)* %out, <2 x half> addrspace(1)*
 
 ; ALL 'fmul_v4f16'
 ; ALL estimated cost of 4 for {{.*}} fmul <4 x half>
-define void @fmul_v4f16(<4 x half> addrspace(1)* %out, <4 x half> addrspace(1)* %vaddr, <4 x half> %b) #0 {
+define amdgpu_kernel void @fmul_v4f16(<4 x half> addrspace(1)* %out, <4 x half> addrspace(1)* %vaddr, <4 x half> %b) #0 {
   %vec = load <4 x half>, <4 x half> addrspace(1)* %vaddr
   %add = fmul <4 x half> %vec, %b
   store <4 x half> %add, <4 x half> addrspace(1)* %out