2 files changed, 18 insertions, 11 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/vgpr-spill-emergency-stack-slot.ll b/llvm/test/CodeGen/AMDGPU/vgpr-spill-emergency-stack-slot.ll
index ed58ce68040..c9c8583d5e8 100644
--- a/llvm/test/CodeGen/AMDGPU/vgpr-spill-emergency-stack-slot.ll
+++ b/llvm/test/CodeGen/AMDGPU/vgpr-spill-emergency-stack-slot.ll
@@ -37,7 +37,8 @@ bb:
   %tmp15 = getelementptr [16 x <16 x i8>], [16 x <16 x i8>] addrspace(2)* %arg4, i64 0, i64 0
   %tmp16 = load <16 x i8>, <16 x i8> addrspace(2)* %tmp15, align 16, !tbaa !0
   %tmp17 = add i32 %arg5, %arg7
-  %tmp18 = call <4 x float> @llvm.SI.vs.load.input(<16 x i8> %tmp16, i32 0, i32 %tmp17)
+  %tmp16.cast = bitcast <16 x i8> %tmp16 to <4 x i32>
+  %tmp18 = call <4 x float> @llvm.amdgcn.buffer.load.format.v4f32(<4 x i32> %tmp16.cast, i32 %tmp17, i32 0, i1 false, i1 false)
   %tmp19 = extractelement <4 x float> %tmp18, i32 0
   %tmp20 = extractelement <4 x float> %tmp18, i32 1
   %tmp21 = extractelement <4 x float> %tmp18, i32 2
@@ -488,10 +489,11 @@ declare i32 @llvm.amdgcn.mbcnt.lo(i32, i32) #1
 declare void @llvm.amdgcn.exp.f32(i32, i32, float, float, float, float, i1, i1) #0
 
 declare float @llvm.SI.load.const(<16 x i8>, i32) #1
-declare <4 x float> @llvm.SI.vs.load.input(<16 x i8>, i32, i32) #1
+declare <4 x float> @llvm.amdgcn.buffer.load.format.v4f32(<4 x i32>, i32, i32, i1, i1) #2
 
 attributes #0 = { nounwind }
 attributes #1 = { nounwind readnone }
+attributes #2 = { nounwind readonly }
 
 !0 = !{!1, !1, i64 0, i32 1}
 !1 = !{!"const", !2}
diff --git a/llvm/test/CodeGen/AMDGPU/wait.ll b/llvm/test/CodeGen/AMDGPU/wait.ll
index 8be5cd92d30..623cbeae8da 100644
--- a/llvm/test/CodeGen/AMDGPU/wait.ll
+++ b/llvm/test/CodeGen/AMDGPU/wait.ll
@@ -15,7 +15,8 @@ define amdgpu_vs void @main(<16 x i8> addrspace(2)* inreg %arg, <16 x i8> addrsp
 main_body:
   %tmp = getelementptr <16 x i8>, <16 x i8> addrspace(2)* %arg3, i32 0
   %tmp10 = load <16 x i8>, <16 x i8> addrspace(2)* %tmp, !tbaa !0
-  %tmp11 = call <4 x float> @llvm.SI.vs.load.input(<16 x i8> %tmp10, i32 0, i32 %arg6)
+  %tmp10.cast = bitcast <16 x i8> %tmp10 to <4 x i32>
+  %tmp11 = call <4 x float> @llvm.amdgcn.buffer.load.format.v4f32(<4 x i32> %tmp10.cast, i32 %arg6, i32 0, i1 false, i1 false)
   %tmp12 = extractelement <4 x float> %tmp11, i32 0
   %tmp13 = extractelement <4 x float> %tmp11, i32 1
   call void @llvm.amdgcn.s.barrier() #1
@@ -23,7 +24,8 @@ main_body:
   %tmp15 = load float, float addrspace(2)* %constptr, align 4
   %tmp16 = getelementptr <16 x i8>, <16 x i8> addrspace(2)* %arg3, i32 1
   %tmp17 = load <16 x i8>, <16 x i8> addrspace(2)* %tmp16, !tbaa !0
-  %tmp18 = call <4 x float> @llvm.SI.vs.load.input(<16 x i8> %tmp17, i32 0, i32 %arg6)
+  %tmp17.cast = bitcast <16 x i8> %tmp17 to <4 x i32>
+  %tmp18 = call <4 x float> @llvm.amdgcn.buffer.load.format.v4f32(<4 x i32> %tmp17.cast, i32 %arg6, i32 0, i1 false, i1 false)
   %tmp19 = extractelement <4 x float> %tmp18, i32 0
   %tmp20 = extractelement <4 x float> %tmp18, i32 1
   %tmp21 = extractelement <4 x float> %tmp18, i32 2
@@ -40,15 +42,17 @@ main_body:
 ; ILPMAX: s_load_dwordx4
 ; ILPMAX: s_waitcnt lgkmcnt(0)
 ; ILPMAX: buffer_load
-; ILPMAX: s_waitcnt vmcnt(1)
 ; ILPMAX: s_waitcnt vmcnt(0)
+; ILPMAX: exp pos0
+; ILPMAX-NEXT: exp param0
 ; ILPMAX: s_endpgm
 define amdgpu_vs void @main2([6 x <16 x i8>] addrspace(2)* byval %arg, [17 x <16 x i8>] addrspace(2)* byval %arg1, [17 x <4 x i32>] addrspace(2)* byval %arg2, [34 x <8 x i32>] addrspace(2)* byval %arg3, [16 x <16 x i8>] addrspace(2)* byval %arg4, i32 inreg %arg5, i32 inreg %arg6, i32 %arg7, i32 %arg8, i32 %arg9, i32 %arg10) #0 {
 main_body:
   %tmp = getelementptr [16 x <16 x i8>], [16 x <16 x i8>] addrspace(2)* %arg4, i64 0, i64 0
   %tmp11 = load <16 x i8>, <16 x i8> addrspace(2)* %tmp, align 16, !tbaa !0
   %tmp12 = add i32 %arg5, %arg7
-  %tmp13 = call <4 x float> @llvm.SI.vs.load.input(<16 x i8> %tmp11, i32 0, i32 %tmp12)
+  %tmp11.cast = bitcast <16 x i8> %tmp11 to <4 x i32>
+  %tmp13 = call <4 x float> @llvm.amdgcn.buffer.load.format.v4f32(<4 x i32> %tmp11.cast, i32 %tmp12, i32 0, i1 false, i1 false)
   %tmp14 = extractelement <4 x float> %tmp13, i32 0
   %tmp15 = extractelement <4 x float> %tmp13, i32 1
   %tmp16 = extractelement <4 x float> %tmp13, i32 2
@@ -56,23 +60,24 @@ main_body:
   %tmp18 = getelementptr [16 x <16 x i8>], [16 x <16 x i8>] addrspace(2)* %arg4, i64 0, i64 1
   %tmp19 = load <16 x i8>, <16 x i8> addrspace(2)* %tmp18, align 16, !tbaa !0
   %tmp20 = add i32 %arg5, %arg7
-  %tmp21 = call <4 x float> @llvm.SI.vs.load.input(<16 x i8> %tmp19, i32 0, i32 %tmp20)
+  %tmp19.cast = bitcast <16 x i8> %tmp19 to <4 x i32>
+  %tmp21 = call <4 x float> @llvm.amdgcn.buffer.load.format.v4f32(<4 x i32> %tmp19.cast, i32 %tmp20, i32 0, i1 false, i1 false)
   %tmp22 = extractelement <4 x float> %tmp21, i32 0
   %tmp23 = extractelement <4 x float> %tmp21, i32 1
   %tmp24 = extractelement <4 x float> %tmp21, i32 2
   %tmp25 = extractelement <4 x float> %tmp21, i32 3
-  call void @llvm.amdgcn.exp.f32(i32 12, i32 15, float %tmp14, float %tmp15, float %tmp16, float %tmp17, i1 true, i1 false) #0
-  call void @llvm.amdgcn.exp.f32(i32 32, i32 15, float %tmp22, float %tmp23, float %tmp24, float %tmp25, i1 false, i1 false) #0
+  call void @llvm.amdgcn.exp.f32(i32 12, i32 15, float %tmp14, float %tmp15, float %tmp16, float %tmp17, i1 false, i1 false) #0
+  call void @llvm.amdgcn.exp.f32(i32 32, i32 15, float %tmp22, float %tmp23, float %tmp24, float %tmp25, i1 true, i1 false) #0
   ret void
 }
 
 declare void @llvm.amdgcn.s.barrier() #1
-declare <4 x float> @llvm.SI.vs.load.input(<16 x i8>, i32, i32) #2
+declare <4 x float> @llvm.amdgcn.buffer.load.format.v4f32(<4 x i32>, i32, i32, i1, i1) #2
 declare void @llvm.amdgcn.exp.f32(i32, i32, float, float, float, float, i1, i1) #0
 
 attributes #0 = { nounwind }
 attributes #1 = { convergent nounwind }
-attributes #2 = { nounwind readnone }
+attributes #2 = { nounwind readonly }
 
 !0 = !{!1, !1, i64 0, i32 1}
 !1 = !{!"const", !2}