AMDGPU: Cleanup load tests
There are a lot of different kinds of loads to test for,
and these were scattered around inconsistently with
some redundancy. Try to comprehensively test all loads
in a consistent way.
llvm-svn: 271571
diff --git a/llvm/test/CodeGen/AMDGPU/extload.ll b/llvm/test/CodeGen/AMDGPU/extload.ll
index b8d7d56..2cb5cf0 100644
--- a/llvm/test/CodeGen/AMDGPU/extload.ll
+++ b/llvm/test/CodeGen/AMDGPU/extload.ll
@@ -1,14 +1,16 @@
-; RUN: llc -march=amdgcn -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
-; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs< %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
-; RUN: llc -march=r600 -mcpu=cypress < %s | FileCheck -check-prefix=EG -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -verify-machineinstrs < %s | FileCheck -check-prefix=SI-NOHSA -check-prefix=FUNC %s
+; RUN: llc -mtriple=amdgcn-amdhsa -mcpu=kaveri -verify-machineinstrs < %s | FileCheck -check-prefix=FUNC -check-prefix=CI-HSA -check-prefix=SI %s
+; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=SI-NOHSA -check-prefix=FUNC %s
+; RUN: llc -march=r600 -mcpu=redwood < %s | FileCheck -check-prefix=EG -check-prefix=FUNC %s
-; FUNC-LABEL: {{^}}anyext_load_i8:
-; SI: buffer_load_dword v{{[0-9]+}}
-; SI: buffer_store_dword v{{[0-9]+}}
+; FIXME: This seems to not ever actually become an extload
+; FUNC-LABEL: {{^}}global_anyext_load_i8:
+; GCN: buffer_load_dword v{{[0-9]+}}
+; GCN: buffer_store_dword v{{[0-9]+}}
; EG: MEM_RAT_CACHELESS STORE_RAW [[VAL:T[0-9]+.[XYZW]]],
; EG: VTX_READ_32 [[VAL]]
-define void @anyext_load_i8(i8 addrspace(1)* nocapture noalias %out, i8 addrspace(1)* nocapture noalias %src) nounwind {
+define void @global_anyext_load_i8(i8 addrspace(1)* nocapture noalias %out, i8 addrspace(1)* nocapture noalias %src) nounwind {
%cast = bitcast i8 addrspace(1)* %src to i32 addrspace(1)*
%load = load i32, i32 addrspace(1)* %cast
%x = bitcast i32 %load to <4 x i8>
@@ -17,13 +19,13 @@
ret void
}
-; FUNC-LABEL: {{^}}anyext_load_i16:
-; SI: buffer_load_dword v{{[0-9]+}}
-; SI: buffer_store_dword v{{[0-9]+}}
+; FUNC-LABEL: {{^}}global_anyext_load_i16:
+; GCN: buffer_load_dword v{{[0-9]+}}
+; GCN: buffer_store_dword v{{[0-9]+}}
; EG: MEM_RAT_CACHELESS STORE_RAW [[VAL:T[0-9]+.[XYZW]]],
; EG: VTX_READ_32 [[VAL]]
-define void @anyext_load_i16(i16 addrspace(1)* nocapture noalias %out, i16 addrspace(1)* nocapture noalias %src) nounwind {
+define void @global_anyext_load_i16(i16 addrspace(1)* nocapture noalias %out, i16 addrspace(1)* nocapture noalias %src) nounwind {
%cast = bitcast i16 addrspace(1)* %src to i32 addrspace(1)*
%load = load i32, i32 addrspace(1)* %cast
%x = bitcast i32 %load to <2 x i16>
@@ -32,13 +34,13 @@
ret void
}
-; FUNC-LABEL: {{^}}anyext_load_lds_i8:
-; SI: ds_read_b32 v{{[0-9]+}}
-; SI: ds_write_b32 v{{[0-9]+}}
+; FUNC-LABEL: {{^}}local_anyext_load_i8:
+; GCN: ds_read_b32 v{{[0-9]+}}
+; GCN: ds_write_b32 v{{[0-9]+}}
; EG: LDS_READ_RET {{.*}}, [[VAL:T[0-9]+.[XYZW]]]
; EG: LDS_WRITE * [[VAL]]
-define void @anyext_load_lds_i8(i8 addrspace(3)* nocapture noalias %out, i8 addrspace(3)* nocapture noalias %src) nounwind {
+define void @local_anyext_load_i8(i8 addrspace(3)* nocapture noalias %out, i8 addrspace(3)* nocapture noalias %src) nounwind {
%cast = bitcast i8 addrspace(3)* %src to i32 addrspace(3)*
%load = load i32, i32 addrspace(3)* %cast
%x = bitcast i32 %load to <4 x i8>
@@ -47,13 +49,13 @@
ret void
}
-; FUNC-LABEL: {{^}}anyext_load_lds_i16:
-; SI: ds_read_b32 v{{[0-9]+}}
-; SI: ds_write_b32 v{{[0-9]+}}
+; FUNC-LABEL: {{^}}local_anyext_load_i16:
+; GCN: ds_read_b32 v{{[0-9]+}}
+; GCN: ds_write_b32 v{{[0-9]+}}
; EG: LDS_READ_RET {{.*}}, [[VAL:T[0-9]+.[XYZW]]]
; EG: LDS_WRITE * [[VAL]]
-define void @anyext_load_lds_i16(i16 addrspace(3)* nocapture noalias %out, i16 addrspace(3)* nocapture noalias %src) nounwind {
+define void @local_anyext_load_i16(i16 addrspace(3)* nocapture noalias %out, i16 addrspace(3)* nocapture noalias %src) nounwind {
%cast = bitcast i16 addrspace(3)* %src to i32 addrspace(3)*
%load = load i32, i32 addrspace(3)* %cast
%x = bitcast i32 %load to <2 x i16>