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/load-constant-i8.ll b/llvm/test/CodeGen/AMDGPU/load-constant-i8.ll
index 9fdc4eb..6e56b9f 100644
--- a/llvm/test/CodeGen/AMDGPU/load-constant-i8.ll
+++ b/llvm/test/CodeGen/AMDGPU/load-constant-i8.ll
@@ -10,7 +10,7 @@
; EG: VTX_READ_8 T{{[0-9]+}}.X, T{{[0-9]+}}.X, 0, #1
; TODO: NOT AND
-define void @constant_load_i8(i8 addrspace(1)* %out, i8 addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_load_i8(i8 addrspace(1)* %out, i8 addrspace(2)* %in) #0 {
entry:
%ld = load i8, i8 addrspace(2)* %in
store i8 %ld, i8 addrspace(1)* %out
@@ -22,7 +22,7 @@
; GCN-HSA: flat_load_ushort v
; EG: VTX_READ_16 T{{[0-9]+}}.X, T{{[0-9]+}}.X, 0, #1
-define void @constant_load_v2i8(<2 x i8> addrspace(1)* %out, <2 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_load_v2i8(<2 x i8> addrspace(1)* %out, <2 x i8> addrspace(2)* %in) #0 {
entry:
%ld = load <2 x i8>, <2 x i8> addrspace(2)* %in
store <2 x i8> %ld, <2 x i8> addrspace(1)* %out
@@ -33,7 +33,7 @@
; GCN: s_load_dword s
; EG: VTX_READ_32 T{{[0-9]+}}.X, T{{[0-9]+}}.X, 0, #1
-define void @constant_load_v3i8(<3 x i8> addrspace(1)* %out, <3 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_load_v3i8(<3 x i8> addrspace(1)* %out, <3 x i8> addrspace(2)* %in) #0 {
entry:
%ld = load <3 x i8>, <3 x i8> addrspace(2)* %in
store <3 x i8> %ld, <3 x i8> addrspace(1)* %out
@@ -44,7 +44,7 @@
; GCN: s_load_dword s
; EG: VTX_READ_32 T{{[0-9]+}}.X, T{{[0-9]+}}.X, 0, #1
-define void @constant_load_v4i8(<4 x i8> addrspace(1)* %out, <4 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_load_v4i8(<4 x i8> addrspace(1)* %out, <4 x i8> addrspace(2)* %in) #0 {
entry:
%ld = load <4 x i8>, <4 x i8> addrspace(2)* %in
store <4 x i8> %ld, <4 x i8> addrspace(1)* %out
@@ -55,7 +55,7 @@
; GCN: s_load_dwordx2
; EG: VTX_READ_64 T{{[0-9]+}}.XY, T{{[0-9]+}}.X, 0, #1
-define void @constant_load_v8i8(<8 x i8> addrspace(1)* %out, <8 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_load_v8i8(<8 x i8> addrspace(1)* %out, <8 x i8> addrspace(2)* %in) #0 {
entry:
%ld = load <8 x i8>, <8 x i8> addrspace(2)* %in
store <8 x i8> %ld, <8 x i8> addrspace(1)* %out
@@ -66,7 +66,7 @@
; GCN: s_load_dwordx4
; EG: VTX_READ_128 T{{[0-9]+}}.XYZW, T{{[0-9]+}}.X, 0, #1
-define void @constant_load_v16i8(<16 x i8> addrspace(1)* %out, <16 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_load_v16i8(<16 x i8> addrspace(1)* %out, <16 x i8> addrspace(2)* %in) #0 {
entry:
%ld = load <16 x i8>, <16 x i8> addrspace(2)* %in
store <16 x i8> %ld, <16 x i8> addrspace(1)* %out
@@ -78,7 +78,7 @@
; GCN-HSA: flat_load_ubyte
; EG: VTX_READ_8 T{{[0-9]+}}.X, T{{[0-9]+}}.X, 0, #1
-define void @constant_zextload_i8_to_i32(i32 addrspace(1)* %out, i8 addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_i8_to_i32(i32 addrspace(1)* %out, i8 addrspace(2)* %in) #0 {
%a = load i8, i8 addrspace(2)* %in
%ext = zext i8 %a to i32
store i32 %ext, i32 addrspace(1)* %out
@@ -92,7 +92,7 @@
; EG: VTX_READ_8 [[DST:T[0-9]+\.X]], T{{[0-9]+}}.X, 0, #1
; EG: BFE_INT {{[* ]*}}T{{[0-9].[XYZW]}}, [[DST]], 0.0, literal
; EG: 8
-define void @constant_sextload_i8_to_i32(i32 addrspace(1)* %out, i8 addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_i8_to_i32(i32 addrspace(1)* %out, i8 addrspace(2)* %in) #0 {
%ld = load i8, i8 addrspace(2)* %in
%ext = sext i8 %ld to i32
store i32 %ext, i32 addrspace(1)* %out
@@ -102,7 +102,7 @@
; FUNC-LABEL: {{^}}constant_zextload_v1i8_to_v1i32:
; EG: VTX_READ_8 T{{[0-9]+}}.X, T{{[0-9]+}}.X, 0, #1
-define void @constant_zextload_v1i8_to_v1i32(<1 x i32> addrspace(1)* %out, <1 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v1i8_to_v1i32(<1 x i32> addrspace(1)* %out, <1 x i8> addrspace(2)* %in) #0 {
%load = load <1 x i8>, <1 x i8> addrspace(2)* %in
%ext = zext <1 x i8> %load to <1 x i32>
store <1 x i32> %ext, <1 x i32> addrspace(1)* %out
@@ -114,7 +114,7 @@
; EG: VTX_READ_8 [[DST:T[0-9]+\.X]], T{{[0-9]+}}.X, 0, #1
; EG: BFE_INT {{[* ]*}}T{{[0-9].[XYZW]}}, [[DST]], 0.0, literal
; EG: 8
-define void @constant_sextload_v1i8_to_v1i32(<1 x i32> addrspace(1)* %out, <1 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v1i8_to_v1i32(<1 x i32> addrspace(1)* %out, <1 x i8> addrspace(2)* %in) #0 {
%load = load <1 x i8>, <1 x i8> addrspace(2)* %in
%ext = sext <1 x i8> %load to <1 x i32>
store <1 x i32> %ext, <1 x i32> addrspace(1)* %out
@@ -129,7 +129,7 @@
; TODO: This should use DST, but for some there are redundant MOVs
; EG: BFE_UINT {{[* ]*}}T{{[0-9].[XYZW]}}, {{.*}}, {{.*}}literal
; EG: 8
-define void @constant_zextload_v2i8_to_v2i32(<2 x i32> addrspace(1)* %out, <2 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v2i8_to_v2i32(<2 x i32> addrspace(1)* %out, <2 x i8> addrspace(2)* %in) #0 {
%load = load <2 x i8>, <2 x i8> addrspace(2)* %in
%ext = zext <2 x i8> %load to <2 x i32>
store <2 x i32> %ext, <2 x i32> addrspace(1)* %out
@@ -150,7 +150,7 @@
; EG-DAG: BFE_INT {{[* ]*}}T{{[0-9].[XYZW]}}, {{.*}}, 0.0, literal
; EG-DAG: 8
; EG-DAG: 8
-define void @constant_sextload_v2i8_to_v2i32(<2 x i32> addrspace(1)* %out, <2 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v2i8_to_v2i32(<2 x i32> addrspace(1)* %out, <2 x i8> addrspace(2)* %in) #0 {
%load = load <2 x i8>, <2 x i8> addrspace(2)* %in
%ext = sext <2 x i8> %load to <2 x i32>
store <2 x i32> %ext, <2 x i32> addrspace(1)* %out
@@ -170,7 +170,7 @@
; EG-DAG: BFE_UINT {{[* ]*}}T{{[0-9].[XYZW]}}, {{.*}}, {{.*}}literal
; EG-DAG: 8
; EG-DAG: 8
-define void @constant_zextload_v3i8_to_v3i32(<3 x i32> addrspace(1)* %out, <3 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v3i8_to_v3i32(<3 x i32> addrspace(1)* %out, <3 x i8> addrspace(2)* %in) #0 {
entry:
%ld = load <3 x i8>, <3 x i8> addrspace(2)* %in
%ext = zext <3 x i8> %ld to <3 x i32>
@@ -193,7 +193,7 @@
; EG-DAG: 8
; EG-DAG: 8
; EG-DAG: 8
-define void @constant_sextload_v3i8_to_v3i32(<3 x i32> addrspace(1)* %out, <3 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v3i8_to_v3i32(<3 x i32> addrspace(1)* %out, <3 x i8> addrspace(2)* %in) #0 {
entry:
%ld = load <3 x i8>, <3 x i8> addrspace(2)* %in
%ext = sext <3 x i8> %ld to <3 x i32>
@@ -214,7 +214,7 @@
; EG-DAG: 8
; EG-DAG: 8
; EG-DAG: 8
-define void @constant_zextload_v4i8_to_v4i32(<4 x i32> addrspace(1)* %out, <4 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v4i8_to_v4i32(<4 x i32> addrspace(1)* %out, <4 x i8> addrspace(2)* %in) #0 {
%load = load <4 x i8>, <4 x i8> addrspace(2)* %in
%ext = zext <4 x i8> %load to <4 x i32>
store <4 x i32> %ext, <4 x i32> addrspace(1)* %out
@@ -236,7 +236,7 @@
; EG-DAG: 8
; EG-DAG: 8
; EG-DAG: 8
-define void @constant_sextload_v4i8_to_v4i32(<4 x i32> addrspace(1)* %out, <4 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v4i8_to_v4i32(<4 x i32> addrspace(1)* %out, <4 x i8> addrspace(2)* %in) #0 {
%load = load <4 x i8>, <4 x i8> addrspace(2)* %in
%ext = sext <4 x i8> %load to <4 x i32>
store <4 x i32> %ext, <4 x i32> addrspace(1)* %out
@@ -264,7 +264,7 @@
; EG-DAG: 8
; EG-DAG: 8
; EG-DAG: 8
-define void @constant_zextload_v8i8_to_v8i32(<8 x i32> addrspace(1)* %out, <8 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v8i8_to_v8i32(<8 x i32> addrspace(1)* %out, <8 x i8> addrspace(2)* %in) #0 {
%load = load <8 x i8>, <8 x i8> addrspace(2)* %in
%ext = zext <8 x i8> %load to <8 x i32>
store <8 x i32> %ext, <8 x i32> addrspace(1)* %out
@@ -294,7 +294,7 @@
; EG-DAG: 8
; EG-DAG: 8
; EG-DAG: 8
-define void @constant_sextload_v8i8_to_v8i32(<8 x i32> addrspace(1)* %out, <8 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v8i8_to_v8i32(<8 x i32> addrspace(1)* %out, <8 x i8> addrspace(2)* %in) #0 {
%load = load <8 x i8>, <8 x i8> addrspace(2)* %in
%ext = sext <8 x i8> %load to <8 x i32>
store <8 x i32> %ext, <8 x i32> addrspace(1)* %out
@@ -335,7 +335,7 @@
; EG-DAG: 8
; EG-DAG: 8
; EG-DAG: 8
-define void @constant_zextload_v16i8_to_v16i32(<16 x i32> addrspace(1)* %out, <16 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v16i8_to_v16i32(<16 x i32> addrspace(1)* %out, <16 x i8> addrspace(2)* %in) #0 {
%load = load <16 x i8>, <16 x i8> addrspace(2)* %in
%ext = zext <16 x i8> %load to <16 x i32>
store <16 x i32> %ext, <16 x i32> addrspace(1)* %out
@@ -378,7 +378,7 @@
; EG-DAG: 8
; EG-DAG: 8
; EG-DAG: 8
-define void @constant_sextload_v16i8_to_v16i32(<16 x i32> addrspace(1)* %out, <16 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v16i8_to_v16i32(<16 x i32> addrspace(1)* %out, <16 x i8> addrspace(2)* %in) #0 {
%load = load <16 x i8>, <16 x i8> addrspace(2)* %in
%ext = sext <16 x i8> %load to <16 x i32>
store <16 x i32> %ext, <16 x i32> addrspace(1)* %out
@@ -450,7 +450,7 @@
; EG-DAG: 8
; EG-DAG: 8
; EG-DAG: 8
-define void @constant_zextload_v32i8_to_v32i32(<32 x i32> addrspace(1)* %out, <32 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v32i8_to_v32i32(<32 x i32> addrspace(1)* %out, <32 x i8> addrspace(2)* %in) #0 {
%load = load <32 x i8>, <32 x i8> addrspace(2)* %in
%ext = zext <32 x i8> %load to <32 x i32>
store <32 x i32> %ext, <32 x i32> addrspace(1)* %out
@@ -526,7 +526,7 @@
; EG-DAG: 8
; EG-DAG: 8
; EG-DAG: 8
-define void @constant_sextload_v32i8_to_v32i32(<32 x i32> addrspace(1)* %out, <32 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v32i8_to_v32i32(<32 x i32> addrspace(1)* %out, <32 x i8> addrspace(2)* %in) #0 {
%load = load <32 x i8>, <32 x i8> addrspace(2)* %in
%ext = sext <32 x i8> %load to <32 x i32>
store <32 x i32> %ext, <32 x i32> addrspace(1)* %out
@@ -539,7 +539,7 @@
; EG-DAG: VTX_READ_128 {{T[0-9]+\.XYZW}}, T{{[0-9]+}}.X, 16, #1
; EG-DAG: VTX_READ_128 {{T[0-9]+\.XYZW}}, T{{[0-9]+}}.X, 32, #1
; EG-DAG: VTX_READ_128 {{T[0-9]+\.XYZW}}, T{{[0-9]+}}.X, 48, #1
-define void @constant_zextload_v64i8_to_v64i32(<64 x i32> addrspace(1)* %out, <64 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v64i8_to_v64i32(<64 x i32> addrspace(1)* %out, <64 x i8> addrspace(2)* %in) #0 {
%load = load <64 x i8>, <64 x i8> addrspace(2)* %in
%ext = zext <64 x i8> %load to <64 x i32>
store <64 x i32> %ext, <64 x i32> addrspace(1)* %out
@@ -552,7 +552,7 @@
; EG-DAG: VTX_READ_128 {{T[0-9]+\.XYZW}}, T{{[0-9]+}}.X, 16, #1
; EG-DAG: VTX_READ_128 {{T[0-9]+\.XYZW}}, T{{[0-9]+}}.X, 32, #1
; EG-DAG: VTX_READ_128 {{T[0-9]+\.XYZW}}, T{{[0-9]+}}.X, 48, #1
-define void @constant_sextload_v64i8_to_v64i32(<64 x i32> addrspace(1)* %out, <64 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v64i8_to_v64i32(<64 x i32> addrspace(1)* %out, <64 x i8> addrspace(2)* %in) #0 {
%load = load <64 x i8>, <64 x i8> addrspace(2)* %in
%ext = sext <64 x i8> %load to <64 x i32>
store <64 x i32> %ext, <64 x i32> addrspace(1)* %out
@@ -570,7 +570,7 @@
; EG: VTX_READ_8 T{{[0-9]+}}.X, T{{[0-9]+}}.X, 0, #1
; EG: MOV {{.*}}, 0.0
-define void @constant_zextload_i8_to_i64(i64 addrspace(1)* %out, i8 addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_i8_to_i64(i64 addrspace(1)* %out, i8 addrspace(2)* %in) #0 {
%a = load i8, i8 addrspace(2)* %in
%ext = zext i8 %a to i64
store i64 %ext, i64 addrspace(1)* %out
@@ -589,7 +589,7 @@
; EG: ASHR {{\**}} {{T[0-9]\.[XYZW]}}, {{.*}}, literal
; TODO: Why not 7 ?
; EG: 31
-define void @constant_sextload_i8_to_i64(i64 addrspace(1)* %out, i8 addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_i8_to_i64(i64 addrspace(1)* %out, i8 addrspace(2)* %in) #0 {
%a = load i8, i8 addrspace(2)* %in
%ext = sext i8 %a to i64
store i64 %ext, i64 addrspace(1)* %out
@@ -600,7 +600,7 @@
; EG: VTX_READ_8 T{{[0-9]+}}.X, T{{[0-9]+}}.X, 0, #1
; EG: MOV {{.*}}, 0.0
-define void @constant_zextload_v1i8_to_v1i64(<1 x i64> addrspace(1)* %out, <1 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v1i8_to_v1i64(<1 x i64> addrspace(1)* %out, <1 x i8> addrspace(2)* %in) #0 {
%load = load <1 x i8>, <1 x i8> addrspace(2)* %in
%ext = zext <1 x i8> %load to <1 x i64>
store <1 x i64> %ext, <1 x i64> addrspace(1)* %out
@@ -613,7 +613,7 @@
; EG: ASHR {{\**}} {{T[0-9]\.[XYZW]}}, {{.*}}, literal
; TODO: Why not 7 ?
; EG: 31
-define void @constant_sextload_v1i8_to_v1i64(<1 x i64> addrspace(1)* %out, <1 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v1i8_to_v1i64(<1 x i64> addrspace(1)* %out, <1 x i8> addrspace(2)* %in) #0 {
%load = load <1 x i8>, <1 x i8> addrspace(2)* %in
%ext = sext <1 x i8> %load to <1 x i64>
store <1 x i64> %ext, <1 x i64> addrspace(1)* %out
@@ -623,7 +623,7 @@
; FUNC-LABEL: {{^}}constant_zextload_v2i8_to_v2i64:
; EG: VTX_READ_16 T{{[0-9]+}}.X, T{{[0-9]+}}.X, 0, #1
-define void @constant_zextload_v2i8_to_v2i64(<2 x i64> addrspace(1)* %out, <2 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v2i8_to_v2i64(<2 x i64> addrspace(1)* %out, <2 x i8> addrspace(2)* %in) #0 {
%load = load <2 x i8>, <2 x i8> addrspace(2)* %in
%ext = zext <2 x i8> %load to <2 x i64>
store <2 x i64> %ext, <2 x i64> addrspace(1)* %out
@@ -633,7 +633,7 @@
; FUNC-LABEL: {{^}}constant_sextload_v2i8_to_v2i64:
; EG: VTX_READ_16 T{{[0-9]+}}.X, T{{[0-9]+}}.X, 0, #1
-define void @constant_sextload_v2i8_to_v2i64(<2 x i64> addrspace(1)* %out, <2 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v2i8_to_v2i64(<2 x i64> addrspace(1)* %out, <2 x i8> addrspace(2)* %in) #0 {
%load = load <2 x i8>, <2 x i8> addrspace(2)* %in
%ext = sext <2 x i8> %load to <2 x i64>
store <2 x i64> %ext, <2 x i64> addrspace(1)* %out
@@ -643,7 +643,7 @@
; FUNC-LABEL: {{^}}constant_zextload_v4i8_to_v4i64:
; EG: VTX_READ_32 T{{[0-9]+}}.X, T{{[0-9]+}}.X, 0, #1
-define void @constant_zextload_v4i8_to_v4i64(<4 x i64> addrspace(1)* %out, <4 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v4i8_to_v4i64(<4 x i64> addrspace(1)* %out, <4 x i8> addrspace(2)* %in) #0 {
%load = load <4 x i8>, <4 x i8> addrspace(2)* %in
%ext = zext <4 x i8> %load to <4 x i64>
store <4 x i64> %ext, <4 x i64> addrspace(1)* %out
@@ -653,7 +653,7 @@
; FUNC-LABEL: {{^}}constant_sextload_v4i8_to_v4i64:
; EG: VTX_READ_32 T{{[0-9]+}}.X, T{{[0-9]+}}.X, 0, #1
-define void @constant_sextload_v4i8_to_v4i64(<4 x i64> addrspace(1)* %out, <4 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v4i8_to_v4i64(<4 x i64> addrspace(1)* %out, <4 x i8> addrspace(2)* %in) #0 {
%load = load <4 x i8>, <4 x i8> addrspace(2)* %in
%ext = sext <4 x i8> %load to <4 x i64>
store <4 x i64> %ext, <4 x i64> addrspace(1)* %out
@@ -663,7 +663,7 @@
; FUNC-LABEL: {{^}}constant_zextload_v8i8_to_v8i64:
; EG: VTX_READ_64 T{{[0-9]+}}.XY, T{{[0-9]+}}.X, 0, #1
-define void @constant_zextload_v8i8_to_v8i64(<8 x i64> addrspace(1)* %out, <8 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v8i8_to_v8i64(<8 x i64> addrspace(1)* %out, <8 x i8> addrspace(2)* %in) #0 {
%load = load <8 x i8>, <8 x i8> addrspace(2)* %in
%ext = zext <8 x i8> %load to <8 x i64>
store <8 x i64> %ext, <8 x i64> addrspace(1)* %out
@@ -673,7 +673,7 @@
; FUNC-LABEL: {{^}}constant_sextload_v8i8_to_v8i64:
; EG: VTX_READ_64 T{{[0-9]+}}.XY, T{{[0-9]+}}.X, 0, #1
-define void @constant_sextload_v8i8_to_v8i64(<8 x i64> addrspace(1)* %out, <8 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v8i8_to_v8i64(<8 x i64> addrspace(1)* %out, <8 x i8> addrspace(2)* %in) #0 {
%load = load <8 x i8>, <8 x i8> addrspace(2)* %in
%ext = sext <8 x i8> %load to <8 x i64>
store <8 x i64> %ext, <8 x i64> addrspace(1)* %out
@@ -683,7 +683,7 @@
; FUNC-LABEL: {{^}}constant_zextload_v16i8_to_v16i64:
; EG: VTX_READ_128 T{{[0-9]+}}.XYZW, T{{[0-9]+}}.X, 0, #1
-define void @constant_zextload_v16i8_to_v16i64(<16 x i64> addrspace(1)* %out, <16 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v16i8_to_v16i64(<16 x i64> addrspace(1)* %out, <16 x i8> addrspace(2)* %in) #0 {
%load = load <16 x i8>, <16 x i8> addrspace(2)* %in
%ext = zext <16 x i8> %load to <16 x i64>
store <16 x i64> %ext, <16 x i64> addrspace(1)* %out
@@ -693,7 +693,7 @@
; FUNC-LABEL: {{^}}constant_sextload_v16i8_to_v16i64:
; EG: VTX_READ_128 T{{[0-9]+}}.XYZW, T{{[0-9]+}}.X, 0, #1
-define void @constant_sextload_v16i8_to_v16i64(<16 x i64> addrspace(1)* %out, <16 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v16i8_to_v16i64(<16 x i64> addrspace(1)* %out, <16 x i8> addrspace(2)* %in) #0 {
%load = load <16 x i8>, <16 x i8> addrspace(2)* %in
%ext = sext <16 x i8> %load to <16 x i64>
store <16 x i64> %ext, <16 x i64> addrspace(1)* %out
@@ -704,7 +704,7 @@
; EG-DAG: VTX_READ_128 T{{[0-9]+}}.XYZW, T{{[0-9]+}}.X, 0, #1
; EG-DAG: VTX_READ_128 T{{[0-9]+}}.XYZW, T{{[0-9]+}}.X, 16, #1
-define void @constant_zextload_v32i8_to_v32i64(<32 x i64> addrspace(1)* %out, <32 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v32i8_to_v32i64(<32 x i64> addrspace(1)* %out, <32 x i8> addrspace(2)* %in) #0 {
%load = load <32 x i8>, <32 x i8> addrspace(2)* %in
%ext = zext <32 x i8> %load to <32 x i64>
store <32 x i64> %ext, <32 x i64> addrspace(1)* %out
@@ -715,7 +715,7 @@
; EG-DAG: VTX_READ_128 T{{[0-9]+}}.XYZW, T{{[0-9]+}}.X, 0, #1
; EG-DAG: VTX_READ_128 T{{[0-9]+}}.XYZW, T{{[0-9]+}}.X, 16, #1
-define void @constant_sextload_v32i8_to_v32i64(<32 x i64> addrspace(1)* %out, <32 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v32i8_to_v32i64(<32 x i64> addrspace(1)* %out, <32 x i8> addrspace(2)* %in) #0 {
%load = load <32 x i8>, <32 x i8> addrspace(2)* %in
%ext = sext <32 x i8> %load to <32 x i64>
store <32 x i64> %ext, <32 x i64> addrspace(1)* %out
@@ -723,7 +723,7 @@
}
; XFUNC-LABEL: {{^}}constant_zextload_v64i8_to_v64i64:
-; define void @constant_zextload_v64i8_to_v64i64(<64 x i64> addrspace(1)* %out, <64 x i8> addrspace(2)* %in) #0 {
+; define amdgpu_kernel void @constant_zextload_v64i8_to_v64i64(<64 x i64> addrspace(1)* %out, <64 x i8> addrspace(2)* %in) #0 {
; %load = load <64 x i8>, <64 x i8> addrspace(2)* %in
; %ext = zext <64 x i8> %load to <64 x i64>
; store <64 x i64> %ext, <64 x i64> addrspace(1)* %out
@@ -731,7 +731,7 @@
; }
; XFUNC-LABEL: {{^}}constant_sextload_v64i8_to_v64i64:
-; define void @constant_sextload_v64i8_to_v64i64(<64 x i64> addrspace(1)* %out, <64 x i8> addrspace(2)* %in) #0 {
+; define amdgpu_kernel void @constant_sextload_v64i8_to_v64i64(<64 x i64> addrspace(1)* %out, <64 x i8> addrspace(2)* %in) #0 {
; %load = load <64 x i8>, <64 x i8> addrspace(2)* %in
; %ext = sext <64 x i8> %load to <64 x i64>
; store <64 x i64> %ext, <64 x i64> addrspace(1)* %out
@@ -744,7 +744,7 @@
; GCN-HSA: flat_load_ubyte v[[VAL:[0-9]+]],
; GCN-HSA: flat_store_short v{{\[[0-9]+:[0-9]+\]}}, v[[VAL]]
-define void @constant_zextload_i8_to_i16(i16 addrspace(1)* %out, i8 addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_i8_to_i16(i16 addrspace(1)* %out, i8 addrspace(2)* %in) #0 {
%a = load i8, i8 addrspace(2)* %in
%ext = zext i8 %a to i16
store i16 %ext, i16 addrspace(1)* %out
@@ -759,7 +759,7 @@
; GCN-HSA: flat_store_short v{{\[[0-9]+:[0-9]+\]}}, v[[VAL]]
; EG: VTX_READ_8 T{{[0-9]+}}.X, T{{[0-9]+}}.X, 0, #1
-define void @constant_sextload_i8_to_i16(i16 addrspace(1)* %out, i8 addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_i8_to_i16(i16 addrspace(1)* %out, i8 addrspace(2)* %in) #0 {
%a = load i8, i8 addrspace(2)* %in
%ext = sext i8 %a to i16
store i16 %ext, i16 addrspace(1)* %out
@@ -767,7 +767,7 @@
}
; FUNC-LABEL: {{^}}constant_zextload_v1i8_to_v1i16:
-define void @constant_zextload_v1i8_to_v1i16(<1 x i16> addrspace(1)* %out, <1 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v1i8_to_v1i16(<1 x i16> addrspace(1)* %out, <1 x i8> addrspace(2)* %in) #0 {
%load = load <1 x i8>, <1 x i8> addrspace(2)* %in
%ext = zext <1 x i8> %load to <1 x i16>
store <1 x i16> %ext, <1 x i16> addrspace(1)* %out
@@ -778,7 +778,7 @@
; EG: VTX_READ_8 T{{[0-9]+}}.X, T{{[0-9]+}}.X, 0, #1
; EG-DAG: BFE_INT {{[* ]*}}T{{[0-9].[XYZW]}}, {{.*}}, 0.0, literal
-define void @constant_sextload_v1i8_to_v1i16(<1 x i16> addrspace(1)* %out, <1 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v1i8_to_v1i16(<1 x i16> addrspace(1)* %out, <1 x i8> addrspace(2)* %in) #0 {
%load = load <1 x i8>, <1 x i8> addrspace(2)* %in
%ext = sext <1 x i8> %load to <1 x i16>
store <1 x i16> %ext, <1 x i16> addrspace(1)* %out
@@ -788,7 +788,7 @@
; FUNC-LABEL: {{^}}constant_zextload_v2i8_to_v2i16:
; EG: VTX_READ_16 T{{[0-9]+}}.X, T{{[0-9]+}}.X, 0, #1
-define void @constant_zextload_v2i8_to_v2i16(<2 x i16> addrspace(1)* %out, <2 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v2i8_to_v2i16(<2 x i16> addrspace(1)* %out, <2 x i8> addrspace(2)* %in) #0 {
%load = load <2 x i8>, <2 x i8> addrspace(2)* %in
%ext = zext <2 x i8> %load to <2 x i16>
store <2 x i16> %ext, <2 x i16> addrspace(1)* %out
@@ -800,7 +800,7 @@
; EG: VTX_READ_16 T{{[0-9]+}}.X, T{{[0-9]+}}.X, 0, #1
; EG-DAG: BFE_INT {{[* ]*}}T{{[0-9].[XYZW]}}, {{.*}}, 0.0, literal
; EG-DAG: BFE_INT {{[* ]*}}T{{[0-9].[XYZW]}}, {{.*}}, 0.0, literal
-define void @constant_sextload_v2i8_to_v2i16(<2 x i16> addrspace(1)* %out, <2 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v2i8_to_v2i16(<2 x i16> addrspace(1)* %out, <2 x i8> addrspace(2)* %in) #0 {
%load = load <2 x i8>, <2 x i8> addrspace(2)* %in
%ext = sext <2 x i8> %load to <2 x i16>
store <2 x i16> %ext, <2 x i16> addrspace(1)* %out
@@ -810,7 +810,7 @@
; FUNC-LABEL: {{^}}constant_zextload_v4i8_to_v4i16:
; EG: VTX_READ_32 T{{[0-9]+}}.X, T{{[0-9]+}}.X, 0, #1
-define void @constant_zextload_v4i8_to_v4i16(<4 x i16> addrspace(1)* %out, <4 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v4i8_to_v4i16(<4 x i16> addrspace(1)* %out, <4 x i8> addrspace(2)* %in) #0 {
%load = load <4 x i8>, <4 x i8> addrspace(2)* %in
%ext = zext <4 x i8> %load to <4 x i16>
store <4 x i16> %ext, <4 x i16> addrspace(1)* %out
@@ -824,7 +824,7 @@
; EG-DAG: BFE_INT {{[* ]*}}T{{[0-9].[XYZW]}}, {{.*}}, 0.0, literal
; EG-DAG: BFE_INT {{[* ]*}}T{{[0-9].[XYZW]}}, {{.*}}, 0.0, literal
; EG-DAG: BFE_INT {{[* ]*}}T{{[0-9].[XYZW]}}, {{.*}}, 0.0, literal
-define void @constant_sextload_v4i8_to_v4i16(<4 x i16> addrspace(1)* %out, <4 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v4i8_to_v4i16(<4 x i16> addrspace(1)* %out, <4 x i8> addrspace(2)* %in) #0 {
%load = load <4 x i8>, <4 x i8> addrspace(2)* %in
%ext = sext <4 x i8> %load to <4 x i16>
store <4 x i16> %ext, <4 x i16> addrspace(1)* %out
@@ -834,7 +834,7 @@
; FUNC-LABEL: {{^}}constant_zextload_v8i8_to_v8i16:
; EG: VTX_READ_64 T{{[0-9]+}}.XY, T{{[0-9]+}}.X, 0, #1
-define void @constant_zextload_v8i8_to_v8i16(<8 x i16> addrspace(1)* %out, <8 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v8i8_to_v8i16(<8 x i16> addrspace(1)* %out, <8 x i8> addrspace(2)* %in) #0 {
%load = load <8 x i8>, <8 x i8> addrspace(2)* %in
%ext = zext <8 x i8> %load to <8 x i16>
store <8 x i16> %ext, <8 x i16> addrspace(1)* %out
@@ -853,7 +853,7 @@
; EG-DAG: BFE_INT {{[* ]*}}T{{[0-9].[XYZW]}}, {{.*}}, 0.0, literal
; EG-DAG: BFE_INT {{[* ]*}}T{{[0-9].[XYZW]}}, {{.*}}, 0.0, literal
-define void @constant_sextload_v8i8_to_v8i16(<8 x i16> addrspace(1)* %out, <8 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v8i8_to_v8i16(<8 x i16> addrspace(1)* %out, <8 x i8> addrspace(2)* %in) #0 {
%load = load <8 x i8>, <8 x i8> addrspace(2)* %in
%ext = sext <8 x i8> %load to <8 x i16>
store <8 x i16> %ext, <8 x i16> addrspace(1)* %out
@@ -863,7 +863,7 @@
; FUNC-LABEL: {{^}}constant_zextload_v16i8_to_v16i16:
; EG: VTX_READ_128 T{{[0-9]+}}.XYZW, T{{[0-9]+}}.X, 0, #1
-define void @constant_zextload_v16i8_to_v16i16(<16 x i16> addrspace(1)* %out, <16 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v16i8_to_v16i16(<16 x i16> addrspace(1)* %out, <16 x i8> addrspace(2)* %in) #0 {
%load = load <16 x i8>, <16 x i8> addrspace(2)* %in
%ext = zext <16 x i8> %load to <16 x i16>
store <16 x i16> %ext, <16 x i16> addrspace(1)* %out
@@ -889,7 +889,7 @@
; EG-DAG: BFE_INT {{[* ]*}}T{{[0-9].[XYZW]}}, {{.*}}, 0.0, literal
; EG-DAG: BFE_INT {{[* ]*}}T{{[0-9].[XYZW]}}, {{.*}}, 0.0, literal
; EG-DAG: BFE_INT {{[* ]*}}T{{[0-9].[XYZW]}}, {{.*}}, 0.0, literal
-define void @constant_sextload_v16i8_to_v16i16(<16 x i16> addrspace(1)* %out, <16 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v16i8_to_v16i16(<16 x i16> addrspace(1)* %out, <16 x i8> addrspace(2)* %in) #0 {
%load = load <16 x i8>, <16 x i8> addrspace(2)* %in
%ext = sext <16 x i8> %load to <16 x i16>
store <16 x i16> %ext, <16 x i16> addrspace(1)* %out
@@ -900,7 +900,7 @@
; EG-DAG: VTX_READ_128 T{{[0-9]+}}.XYZW, T{{[0-9]+}}.X, 0, #1
; EG-DAG: VTX_READ_128 T{{[0-9]+}}.XYZW, T{{[0-9]+}}.X, 16, #1
-define void @constant_zextload_v32i8_to_v32i16(<32 x i16> addrspace(1)* %out, <32 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_zextload_v32i8_to_v32i16(<32 x i16> addrspace(1)* %out, <32 x i8> addrspace(2)* %in) #0 {
%load = load <32 x i8>, <32 x i8> addrspace(2)* %in
%ext = zext <32 x i8> %load to <32 x i16>
store <32 x i16> %ext, <32 x i16> addrspace(1)* %out
@@ -943,7 +943,7 @@
; EG-DAG: BFE_INT {{[* ]*}}T{{[0-9].[XYZW]}}, {{.*}}, 0.0, literal
; EG-DAG: BFE_INT {{[* ]*}}T{{[0-9].[XYZW]}}, {{.*}}, 0.0, literal
; EG-DAG: BFE_INT {{[* ]*}}T{{[0-9].[XYZW]}}, {{.*}}, 0.0, literal
-define void @constant_sextload_v32i8_to_v32i16(<32 x i16> addrspace(1)* %out, <32 x i8> addrspace(2)* %in) #0 {
+define amdgpu_kernel void @constant_sextload_v32i8_to_v32i16(<32 x i16> addrspace(1)* %out, <32 x i8> addrspace(2)* %in) #0 {
%load = load <32 x i8>, <32 x i8> addrspace(2)* %in
%ext = sext <32 x i8> %load to <32 x i16>
store <32 x i16> %ext, <32 x i16> addrspace(1)* %out
@@ -951,7 +951,7 @@
}
; XFUNC-LABEL: {{^}}constant_zextload_v64i8_to_v64i16:
-; define void @constant_zextload_v64i8_to_v64i16(<64 x i16> addrspace(1)* %out, <64 x i8> addrspace(2)* %in) #0 {
+; define amdgpu_kernel void @constant_zextload_v64i8_to_v64i16(<64 x i16> addrspace(1)* %out, <64 x i8> addrspace(2)* %in) #0 {
; %load = load <64 x i8>, <64 x i8> addrspace(2)* %in
; %ext = zext <64 x i8> %load to <64 x i16>
; store <64 x i16> %ext, <64 x i16> addrspace(1)* %out
@@ -959,7 +959,7 @@
; }
; XFUNC-LABEL: {{^}}constant_sextload_v64i8_to_v64i16:
-; define void @constant_sextload_v64i8_to_v64i16(<64 x i16> addrspace(1)* %out, <64 x i8> addrspace(2)* %in) #0 {
+; define amdgpu_kernel void @constant_sextload_v64i8_to_v64i16(<64 x i16> addrspace(1)* %out, <64 x i8> addrspace(2)* %in) #0 {
; %load = load <64 x i8>, <64 x i8> addrspace(2)* %in
; %ext = sext <64 x i8> %load to <64 x i16>
; store <64 x i16> %ext, <64 x i16> addrspace(1)* %out