[AMDGPU][NFC] Update left over tests for COV5 (#76984)

Update AMDGPU CodeGen lit tests to check for COV5 ABI.

13 files changed, 106 insertions, 62 deletions

diff --git a/llvm/test/CodeGen/AMDGPU/attributor-noopt.ll b/llvm/test/CodeGen/AMDGPU/attributor-noopt.ll
index d83884635546..a374689da573 100644
--- a/llvm/test/CodeGen/AMDGPU/attributor-noopt.ll
+++ b/llvm/test/CodeGen/AMDGPU/attributor-noopt.ll
@@ -1,5 +1,6 @@
-; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 < %s | FileCheck -check-prefix=OPT %s
-; RUN: llc -O0 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 < %s | FileCheck -check-prefix=NOOPT %s
+; RUN: sed 's/CODE_OBJECT_VERSION/500/g' %s | llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 | FileCheck -check-prefix=OPT %s
+; RUN: sed 's/CODE_OBJECT_VERSION/400/g' %s | llc -O0 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 | FileCheck -check-prefixes=NOOPT,COV4 %s
+; RUN: sed 's/CODE_OBJECT_VERSION/500/g' %s | llc -O0 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 | FileCheck -check-prefixes=NOOPT,COV5 %s
 
 ; Check that AMDGPUAttributor is not run with -O0.
 ; OPT: .amdhsa_user_sgpr_private_segment_buffer 1
@@ -18,7 +19,8 @@
 
 ; NOOPT: .amdhsa_user_sgpr_private_segment_buffer 1
 ; NOOPT: .amdhsa_user_sgpr_dispatch_ptr 1
-; NOOPT: .amdhsa_user_sgpr_queue_ptr 1
+; COV4: .amdhsa_user_sgpr_queue_ptr 1
+; COV5: .amdhsa_user_sgpr_queue_ptr 0
 ; NOOPT: .amdhsa_user_sgpr_kernarg_segment_ptr 1
 ; NOOPT: .amdhsa_user_sgpr_dispatch_id 1
 ; NOOPT: .amdhsa_user_sgpr_flat_scratch_init 0
@@ -32,3 +34,6 @@
 define amdgpu_kernel void @foo() {
   ret void
 }
+
+!llvm.module.flags = !{!0}
+!0 = !{i32 1, !"amdgpu_code_object_version", i32 CODE_OBJECT_VERSION}
diff --git a/llvm/test/CodeGen/AMDGPU/call-alias-register-usage-agpr.ll b/llvm/test/CodeGen/AMDGPU/call-alias-register-usage-agpr.ll
index 6ff2dbabc8ec..72bb515ba57e 100644
--- a/llvm/test/CodeGen/AMDGPU/call-alias-register-usage-agpr.ll
+++ b/llvm/test/CodeGen/AMDGPU/call-alias-register-usage-agpr.ll
@@ -9,10 +9,10 @@
 
 ; ALL-LABEL: {{^}}kernel:
 ; GFX908: .amdhsa_next_free_vgpr 32
-; GFX908-NEXT: .amdhsa_next_free_sgpr 36
+; GFX908-NEXT: .amdhsa_next_free_sgpr 33
 
-; GFX90A: .amdhsa_next_free_vgpr 65
-; GFX90A-NEXT: .amdhsa_next_free_sgpr 36
+; GFX90A: .amdhsa_next_free_vgpr 59
+; GFX90A-NEXT: .amdhsa_next_free_sgpr 33
 ; GFX90A-NEXT: .amdhsa_accum_offset 32
 define amdgpu_kernel void @kernel() #0 {
 bb:
@@ -29,3 +29,6 @@ bb:
 attributes #0 = { noinline norecurse nounwind optnone }
 attributes #1 = { noinline norecurse nounwind readnone willreturn }
 attributes #2 = { nounwind readnone willreturn }
+
+!llvm.module.flags = !{!0}
+!0 = !{i32 1, !"amdgpu_code_object_version", i32 500}
diff --git a/llvm/test/CodeGen/AMDGPU/call-alias-register-usage0.ll b/llvm/test/CodeGen/AMDGPU/call-alias-register-usage0.ll
index 797376535cde..6afc90639dce 100644
--- a/llvm/test/CodeGen/AMDGPU/call-alias-register-usage0.ll
+++ b/llvm/test/CodeGen/AMDGPU/call-alias-register-usage0.ll
@@ -8,7 +8,7 @@
 
 ; CHECK-LABEL: {{^}}kernel0:
 ; CHECK: .amdhsa_next_free_vgpr 53
-; CHECK-NEXT: .amdhsa_next_free_sgpr 36
+; CHECK-NEXT: .amdhsa_next_free_sgpr 33
 define amdgpu_kernel void @kernel0() #0 {
 bb:
   call void @alias0() #2
@@ -24,3 +24,6 @@ bb:
 attributes #0 = { noinline norecurse nounwind optnone }
 attributes #1 = { noinline norecurse nounwind readnone willreturn }
 attributes #2 = { nounwind readnone willreturn }
+
+!llvm.module.flags = !{!0}
+!0 = !{i32 1, !"amdgpu_code_object_version", i32 500}
diff --git a/llvm/test/CodeGen/AMDGPU/call-alias-register-usage1.ll b/llvm/test/CodeGen/AMDGPU/call-alias-register-usage1.ll
index 79bb2fb64117..137bb13330fd 100644
--- a/llvm/test/CodeGen/AMDGPU/call-alias-register-usage1.ll
+++ b/llvm/test/CodeGen/AMDGPU/call-alias-register-usage1.ll
@@ -10,7 +10,7 @@
 
 ; CHECK-LABEL: {{^}}kernel1:
 ; CHECK: .amdhsa_next_free_vgpr 41
-; CHECK-NEXT: .amdhsa_next_free_sgpr 36
+; CHECK-NEXT: .amdhsa_next_free_sgpr 33
 define amdgpu_kernel void @kernel1() #0 {
 bb:
   call void asm sideeffect "; clobber v40 ", "~{v40}"()
@@ -27,3 +27,6 @@ bb:
 attributes #0 = { noinline norecurse nounwind optnone }
 attributes #1 = { noinline norecurse nounwind readnone willreturn "amdgpu-waves-per-eu"="8,10" }
 attributes #2 = { nounwind readnone willreturn }
+
+!llvm.module.flags = !{!0}
+!0 = !{i32 1, !"amdgpu_code_object_version", i32 500}
diff --git a/llvm/test/CodeGen/AMDGPU/call-alias-register-usage2.ll b/llvm/test/CodeGen/AMDGPU/call-alias-register-usage2.ll
index 5745dd9fdb67..2800ed635bdb 100644
--- a/llvm/test/CodeGen/AMDGPU/call-alias-register-usage2.ll
+++ b/llvm/test/CodeGen/AMDGPU/call-alias-register-usage2.ll
@@ -8,7 +8,7 @@
 
 ; CHECK-LABEL: {{^}}kernel2:
 ; CHECK: .amdhsa_next_free_vgpr 53
-; CHECK-NEXT: .amdhsa_next_free_sgpr 36
+; CHECK-NEXT: .amdhsa_next_free_sgpr 33
 define amdgpu_kernel void @kernel2() #0 {
 bb:
   call void @alias2() #2
@@ -24,3 +24,6 @@ bb:
 attributes #0 = { noinline norecurse nounwind optnone }
 attributes #1 = { noinline norecurse nounwind readnone willreturn "amdgpu-waves-per-eu"="4,10" }
 attributes #2 = { nounwind readnone willreturn }
+
+!llvm.module.flags = !{!0}
+!0 = !{i32 1, !"amdgpu_code_object_version", i32 500}
diff --git a/llvm/test/CodeGen/AMDGPU/call-alias-register-usage3.ll b/llvm/test/CodeGen/AMDGPU/call-alias-register-usage3.ll
index b922297c493f..f7c0a57f5217 100644
--- a/llvm/test/CodeGen/AMDGPU/call-alias-register-usage3.ll
+++ b/llvm/test/CodeGen/AMDGPU/call-alias-register-usage3.ll
@@ -8,7 +8,7 @@
 
 ; CHECK-LABEL: {{^}}kernel3:
 ; CHECK: .amdhsa_next_free_vgpr 253
-; CHECK-NEXT: .amdhsa_next_free_sgpr 36
+; CHECK-NEXT: .amdhsa_next_free_sgpr 33
 define amdgpu_kernel void @kernel3() #0 {
 bb:
   call void @alias3() #2
@@ -24,3 +24,6 @@ bb:
 attributes #0 = { noinline norecurse nounwind optnone }
 attributes #1 = { noinline norecurse nounwind readnone willreturn "amdgpu-flat-work-group-size"="1,256" "amdgpu-waves-per-eu"="1,1" }
 attributes #2 = { nounwind readnone willreturn }
+
+!llvm.module.flags = !{!0}
+!0 = !{i32 1, !"amdgpu_code_object_version", i32 500}
diff --git a/llvm/test/CodeGen/AMDGPU/dagcombine-lshr-and-cmp.ll b/llvm/test/CodeGen/AMDGPU/dagcombine-lshr-and-cmp.ll
index 084b9686f88a..ce478d41380a 100644
--- a/llvm/test/CodeGen/AMDGPU/dagcombine-lshr-and-cmp.ll
+++ b/llvm/test/CodeGen/AMDGPU/dagcombine-lshr-and-cmp.ll
@@ -28,7 +28,6 @@ define i32 @divergent_lshr_and_cmp(i32 %x) {
 entry:
   %0 = and i32 %x, 2
   %1 = icmp ne i32 %0, 0
-  ; Prevent removal of truncate in SDag by inserting llvm.amdgcn.if
   br i1 %1, label %out.true, label %out.else
 
 out.true:
@@ -43,9 +42,9 @@ define amdgpu_kernel void @uniform_opt_lshr_and_cmp(ptr addrspace(1) %out, i32 %
   ; GCN-LABEL: name: uniform_opt_lshr_and_cmp
   ; GCN: bb.0.entry:
   ; GCN-NEXT:   successors: %bb.1(0x40000000), %bb.2(0x40000000)
-  ; GCN-NEXT:   liveins: $sgpr4_sgpr5
+  ; GCN-NEXT:   liveins: $sgpr2_sgpr3
   ; GCN-NEXT: {{  $}}
-  ; GCN-NEXT:   [[COPY:%[0-9]+]]:sgpr_64(p4) = COPY $sgpr4_sgpr5
+  ; GCN-NEXT:   [[COPY:%[0-9]+]]:sgpr_64(p4) = COPY $sgpr2_sgpr3
   ; GCN-NEXT:   [[S_LOAD_DWORDX2_IMM:%[0-9]+]]:sreg_64_xexec = S_LOAD_DWORDX2_IMM [[COPY]](p4), 9, 0 :: (dereferenceable invariant load (s64) from %ir.out.kernarg.offset, align 4, addrspace 4)
   ; GCN-NEXT:   [[S_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_LOAD_DWORD_IMM [[COPY]](p4), 11, 0 :: (dereferenceable invariant load (s32) from %ir.x.kernarg.offset, addrspace 4)
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:sreg_64 = COPY [[S_LOAD_DWORDX2_IMM]]
@@ -84,7 +83,6 @@ define amdgpu_kernel void @uniform_opt_lshr_and_cmp(ptr addrspace(1) %out, i32 %
 entry:
   %0 = and i32 %x, 2
   %1 = icmp ne i32 %0, 0
-  ; Don't optimize the truncate in the SDag away.
   br i1 %1, label %out.true, label %out.else
 
 out.true:
@@ -96,3 +94,6 @@ out.else:
   store i1 %1, ptr addrspace(1) %out
   ret void
 }
+
+!llvm.module.flags = !{!0}
+!0 = !{i32 1, !"amdgpu_code_object_version", i32 500}
diff --git a/llvm/test/CodeGen/AMDGPU/fneg-fabs.ll b/llvm/test/CodeGen/AMDGPU/fneg-fabs.ll
index e2e1effc9b40..3be2d9435119 100644
--- a/llvm/test/CodeGen/AMDGPU/fneg-fabs.ll
+++ b/llvm/test/CodeGen/AMDGPU/fneg-fabs.ll
@@ -49,7 +49,7 @@ define amdgpu_kernel void @fneg_fabsf_free_f32(ptr addrspace(1) %out, i32 %in) {
 ; R600: |PV.{{[XYZW]}}|
 ; R600: -PV
 
-; SI: s_or_b32 s{{[0-9]+}}, s{{[0-9]+}}, 0x80000000
+; SI: s_load_dwordx2 s[0:1], s[2:3], 0x9
 define amdgpu_kernel void @fneg_fabsf_fn_free_f32(ptr addrspace(1) %out, i32 %in) {
   %bc = bitcast i32 %in to float
   %fabs = call float @fabsf(float %bc)
@@ -109,3 +109,6 @@ declare float @fabsf(float) readnone
 declare float @llvm.fabs.f32(float) readnone
 declare <2 x float> @llvm.fabs.v2f32(<2 x float>) readnone
 declare <4 x float> @llvm.fabs.v4f32(<4 x float>) readnone
+
+!llvm.module.flags = !{!0}
+!0 = !{i32 1, !"amdgpu_code_object_version", i32 500}
diff --git a/llvm/test/CodeGen/AMDGPU/gfx11-user-sgpr-init16-bug.ll b/llvm/test/CodeGen/AMDGPU/gfx11-user-sgpr-init16-bug.ll
index ea5add023d15..3973cf1eec83 100644
--- a/llvm/test/CodeGen/AMDGPU/gfx11-user-sgpr-init16-bug.ll
+++ b/llvm/test/CodeGen/AMDGPU/gfx11-user-sgpr-init16-bug.ll
@@ -36,6 +36,7 @@
 ; GCN-NEXT: .amdhsa_user_sgpr_dispatch_id 0
 ; GCN-NEXT: .amdhsa_user_sgpr_private_segment_size 0
 ; GCN-NEXT: .amdhsa_wavefront_size32
+; GCN-NEXT: .amdhsa_uses_dynamic_stack 0
 ; GCN-NEXT: .amdhsa_enable_private_segment 0
 ; GCN-NEXT: .amdhsa_system_sgpr_workgroup_id_x 1
 ; GCN-NEXT: .amdhsa_system_sgpr_workgroup_id_y 0
@@ -64,6 +65,7 @@ define amdgpu_kernel void @minimal_kernel_inputs() {
 ; GCN-NEXT: .amdhsa_user_sgpr_dispatch_id 0
 ; GCN-NEXT: .amdhsa_user_sgpr_private_segment_size 0
 ; GCN-NEXT: .amdhsa_wavefront_size32
+; GCN-NEXT: .amdhsa_uses_dynamic_stack 0
 ; GCN-NEXT: .amdhsa_enable_private_segment 1
 ; GCN-NEXT: .amdhsa_system_sgpr_workgroup_id_x 1
 ; GCN-NEXT: .amdhsa_system_sgpr_workgroup_id_y 0
@@ -81,7 +83,7 @@ define amdgpu_kernel void @minimal_kernel_inputs_with_stack() {
 }
 
 ; GCN-LABEL: {{^}}queue_ptr:
-; GCN: global_load_u8 v{{[0-9]+}}, v{{[0-9]+}}, s[0:1]
+; GCN: global_load_u8 v{{[0-9]+}},
 
 ; WORKAROUND: v_mov_b32_e32 [[V:v[0-9]+]], s15
 ; NOWORKAROUND: v_mov_b32_e32 [[V:v[0-9]+]], s2
@@ -91,11 +93,12 @@ define amdgpu_kernel void @minimal_kernel_inputs_with_stack() {
 ; WORKAROUND: .amdhsa_user_sgpr_count 15
 ; NOWORKAROUND: .amdhsa_user_sgpr_count 2
 ; GCN-NEXT: .amdhsa_user_sgpr_dispatch_ptr 0
-; GCN-NEXT: .amdhsa_user_sgpr_queue_ptr 1
-; GCN-NEXT: .amdhsa_user_sgpr_kernarg_segment_ptr 0
+; GCN-NEXT: .amdhsa_user_sgpr_queue_ptr 0
+; GCN-NEXT: .amdhsa_user_sgpr_kernarg_segment_ptr 1
 ; GCN-NEXT: .amdhsa_user_sgpr_dispatch_id 0
 ; GCN-NEXT: .amdhsa_user_sgpr_private_segment_size 0
 ; GCN-NEXT: .amdhsa_wavefront_size32
+; GCN-NEXT: .amdhsa_uses_dynamic_stack 0
 ; GCN-NEXT: .amdhsa_enable_private_segment 0
 ; GCN-NEXT: .amdhsa_system_sgpr_workgroup_id_x 1
 ; GCN-NEXT: .amdhsa_system_sgpr_workgroup_id_y 0
@@ -117,16 +120,16 @@ define amdgpu_kernel void @queue_ptr() {
 ; WORKAROUND: v_mov_b32_e32 [[V_Y:v[0-9]+]], s14
 ; WORKAROUND: v_mov_b32_e32 [[V_Z:v[0-9]+]], s15
 
-; NOWORKAROUND: v_mov_b32_e32 [[V_X:v[0-9]+]], s8
-; NOWORKAROUND: v_mov_b32_e32 [[V_Y:v[0-9]+]], s9
-; NOWORKAROUND: v_mov_b32_e32 [[V_Z:v[0-9]+]], s10
+; NOWORKAROUND: v_mov_b32_e32 [[V_X:v[0-9]+]], s6
+; NOWORKAROUND: v_mov_b32_e32 [[V_Y:v[0-9]+]], s7
+; NOWORKAROUND: v_mov_b32_e32 [[V_Z:v[0-9]+]], s8
 
 ; GCN: global_load_u8 v{{[0-9]+}}, v{{[0-9]+}}, s[0:1]
+; GCN: global_load_u8 v{{[0-9]+}},
 ; GCN: global_load_u8 v{{[0-9]+}}, v{{[0-9]+}}, s[2:3]
-; GCN: global_load_u8 v{{[0-9]+}}, v{{[0-9]+}}, s[4:5]
 
-; GCN-DAG: v_mov_b32_e32 v[[DISPATCH_LO:[0-9]+]], s6
-; GCN-DAG: v_mov_b32_e32 v[[DISPATCH_HI:[0-9]+]], s7
+; GCN-DAG: v_mov_b32_e32 v[[DISPATCH_LO:[0-9]+]], s4
+; GCN-DAG: v_mov_b32_e32 v[[DISPATCH_HI:[0-9]+]], s5
 
 ; GCN: global_store_b32 v{{\[[0-9]+:[0-9]+\]}}, [[V_X]], off
 ; GCN: global_store_b32 v{{\[[0-9]+:[0-9]+\]}}, [[V_Y]], off
@@ -135,13 +138,14 @@ define amdgpu_kernel void @queue_ptr() {
 
 ; GCN: .amdhsa_kernel all_inputs
 ; WORKAROUND: .amdhsa_user_sgpr_count 13
-; NOWORKAROUND: .amdhsa_user_sgpr_count 8
+; NOWORKAROUND: .amdhsa_user_sgpr_count 6
 ; GCN-NEXT: .amdhsa_user_sgpr_dispatch_ptr 1
-; GCN-NEXT: .amdhsa_user_sgpr_queue_ptr 1
+; GCN-NEXT: .amdhsa_user_sgpr_queue_ptr 0
 ; GCN-NEXT: .amdhsa_user_sgpr_kernarg_segment_ptr 1
 ; GCN-NEXT: .amdhsa_user_sgpr_dispatch_id 1
 ; GCN-NEXT: .amdhsa_user_sgpr_private_segment_size 0
 ; GCN-NEXT: .amdhsa_wavefront_size32
+; GCN-NEXT: .amdhsa_uses_dynamic_stack 0
 ; GCN-NEXT: .amdhsa_enable_private_segment 1
 ; GCN-NEXT: .amdhsa_system_sgpr_workgroup_id_x 1
 ; GCN-NEXT: .amdhsa_system_sgpr_workgroup_id_y 1
@@ -149,7 +153,7 @@ define amdgpu_kernel void @queue_ptr() {
 ; GCN-NEXT: .amdhsa_system_sgpr_workgroup_info 0
 ; GCN-NEXT: .amdhsa_system_vgpr_workitem_id 0
 ; WORKAROUND: ; COMPUTE_PGM_RSRC2:USER_SGPR: 13
-; NOWORKAROUND: ; COMPUTE_PGM_RSRC2:USER_SGPR: 8
+; NOWORKAROUND: ; COMPUTE_PGM_RSRC2:USER_SGPR: 6
 define amdgpu_kernel void @all_inputs() {
   %alloca = alloca i32, addrspace(5)
   store volatile i32 0, ptr addrspace(5) %alloca
@@ -188,3 +192,6 @@ declare align 4 ptr addrspace(4) @llvm.amdgcn.kernarg.segment.ptr() #0
 declare i64 @llvm.amdgcn.dispatch.id() #0
 
 attributes #0 = { nounwind readnone speculatable willreturn }
+
+!llvm.module.flags = !{!0}
+!0 = !{i32 1, !"amdgpu_code_object_version", i32 500}
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.is.shared.ll b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.is.shared.ll
index 7479fc87e318..2672c12ecf1f 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.is.shared.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.is.shared.ll
@@ -1,15 +1,16 @@
-; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=tahiti -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,CI %s
-; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=hawaii -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,CI %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=tahiti -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,CI,CIT %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=hawaii -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,CI,CIH %s
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,GFX9 %s
 
 ; GCN-LABEL: {{^}}is_local_vgpr:
 ; GCN-DAG: {{flat|global|buffer}}_load_dwordx2 v{{\[[0-9]+}}:[[PTR_HI:[0-9]+]]]
-; CI-DAG: s_load_dword [[APERTURE:s[0-9]+]], s[4:5], 0x10
+; CI-DAG: s_load_dwordx2 s[0:1], s[4:5], 0x0
 
 ; GFX9: s_mov_b64 s[{{[0-9]+}}:[[HI:[0-9]+]]], src_shared_base
 ; GFX9: v_cmp_eq_u32_e32 vcc, s[[HI]], v[[PTR_HI]]
 
-; CI: v_cmp_eq_u32_e32 vcc, [[APERTURE]], v[[PTR_HI]]
+; CIT: v_cmp_eq_u32_e32 vcc, s4, v[[PTR_HI]]
+; CIH: v_cmp_eq_u32_e32 vcc, s2, v[[PTR_HI]]
 ; GCN: v_cndmask_b32_e64 v{{[0-9]+}}, 0, 1, vcc
 define amdgpu_kernel void @is_local_vgpr(ptr addrspace(1) %ptr.ptr) {
   %id = call i32 @llvm.amdgcn.workitem.id.x()
@@ -25,15 +26,15 @@ define amdgpu_kernel void @is_local_vgpr(ptr addrspace(1) %ptr.ptr) {
 ; select and vcc branch.
 
 ; GCN-LABEL: {{^}}is_local_sgpr:
-; CI-DAG: s_load_dword [[APERTURE:s[0-9]+]], s[4:5], 0x10{{$}}
+; CI-DAG: s_load_dword s0, s[4:5], 0x1
 
-; CI-DAG: s_load_dword [[PTR_HI:s[0-9]+]], s[6:7], 0x1{{$}}
+; CI-DAG: s_load_dword [[PTR_HI:s[0-9]+]], s[4:5], 0x33{{$}}
 ; GFX9-DAG: s_load_dword [[PTR_HI:s[0-9]+]], s[4:5], 0x4{{$}}
 
 ; GFX9: s_mov_b64 s[{{[0-9]+}}:[[HI:[0-9]+]]], src_shared_base
 ; GFX9: s_cmp_eq_u32 [[PTR_HI]], s[[HI]]
 
-; CI: s_cmp_eq_u32 [[PTR_HI]], [[APERTURE]]
+; CI: s_cmp_eq_u32 s0, [[PTR_HI]]
 ; GCN: s_cbranch_vccnz
 define amdgpu_kernel void @is_local_sgpr(ptr %ptr) {
   %val = call i1 @llvm.amdgcn.is.shared(ptr %ptr)
@@ -51,3 +52,6 @@ declare i32 @llvm.amdgcn.workitem.id.x() #0
 declare i1 @llvm.amdgcn.is.shared(ptr nocapture) #0
 
 attributes #0 = { nounwind readnone speculatable }
+
+!llvm.module.flags = !{!0}
+!0 = !{i32 1, !"amdgpu_code_object_version", i32 500}
diff --git a/llvm/test/CodeGen/AMDGPU/promote-alloca-calling-conv.ll b/llvm/test/CodeGen/AMDGPU/promote-alloca-calling-conv.ll
index ec83d7f313d6..a8bb36bc0d19 100644
--- a/llvm/test/CodeGen/AMDGPU/promote-alloca-calling-conv.ll
+++ b/llvm/test/CodeGen/AMDGPU/promote-alloca-calling-conv.ll
@@ -77,7 +77,7 @@ declare i32 @foo(ptr addrspace(5)) #0
 ; ASM: buffer_store_dword
 ; ASM: buffer_store_dword
 ; ASM: s_swappc_b64
-; ASM: ScratchSize: 16400
+; ASM: ScratchSize: 16
 define amdgpu_kernel void @call_private(ptr addrspace(1) %out, i32 %in) #0 {
 entry:
   %tmp = alloca [2 x i32], addrspace(5)
@@ -94,3 +94,6 @@ declare i32 @llvm.amdgcn.workitem.id.x() #1
 
 attributes #0 = { nounwind "amdgpu-flat-work-group-size"="64,64" }
 attributes #1 = { nounwind readnone }
+
+!llvm.module.flags = !{!0}
+!0 = !{i32 1, !"amdgpu_code_object_version", i32 500}
diff --git a/llvm/test/CodeGen/AMDGPU/reqd-work-group-size.ll b/llvm/test/CodeGen/AMDGPU/reqd-work-group-size.ll
index ecdc3845efc4..7d7917e0b20c 100644
--- a/llvm/test/CodeGen/AMDGPU/reqd-work-group-size.ll
+++ b/llvm/test/CodeGen/AMDGPU/reqd-work-group-size.ll
@@ -24,7 +24,7 @@ define amdgpu_kernel void @volatile_load_group_size_x(ptr addrspace(1) %out) #0
 }
 
 ; CHECK-LABEL: @load_group_size_x(
-; CHECK-NEXT: store i16 8,
+; CHECK: store i16 %group.size.x,
 define amdgpu_kernel void @load_group_size_x(ptr addrspace(1) %out) #0 !reqd_work_group_size !0 {
   %dispatch.ptr = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr()
   %gep.group.size.x = getelementptr inbounds i8, ptr addrspace(4) %dispatch.ptr, i64 4
@@ -34,7 +34,7 @@ define amdgpu_kernel void @load_group_size_x(ptr addrspace(1) %out) #0 !reqd_wor
 }
 
 ; CHECK-LABEL: @load_group_size_y(
-; CHECK-NEXT: store i16 16,
+; CHECK: store i16 %group.size.y,
 define amdgpu_kernel void @load_group_size_y(ptr addrspace(1) %out) #0 !reqd_work_group_size !0 {
   %dispatch.ptr = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr()
   %gep.group.size.y = getelementptr inbounds i8, ptr addrspace(4) %dispatch.ptr, i64 6
@@ -44,7 +44,7 @@ define amdgpu_kernel void @load_group_size_y(ptr addrspace(1) %out) #0 !reqd_wor
 }
 
 ; CHECK-LABEL: @load_group_size_z(
-; CHECK-NEXT: store i16 2,
+; CHECK: store i16 %group.size.z,
 define amdgpu_kernel void @load_group_size_z(ptr addrspace(1) %out) #0 !reqd_work_group_size !0 {
   %dispatch.ptr = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr()
   %gep.group.size.z = getelementptr inbounds i8, ptr addrspace(4) %dispatch.ptr, i64 8
@@ -55,7 +55,7 @@ define amdgpu_kernel void @load_group_size_z(ptr addrspace(1) %out) #0 !reqd_wor
 
 ; Metadata uses i64 instead of i32
 ; CHECK-LABEL: @load_group_size_x_reqd_work_group_size_i64(
-; CHECK-NEXT: store i16 8,
+; CHECK: store i16 %group.size.x,
 define amdgpu_kernel void @load_group_size_x_reqd_work_group_size_i64(ptr addrspace(1) %out) #0 !reqd_work_group_size !2 {
   %dispatch.ptr = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr()
   %gep.group.size.x = getelementptr inbounds i8, ptr addrspace(4) %dispatch.ptr, i64 4
@@ -66,7 +66,7 @@ define amdgpu_kernel void @load_group_size_x_reqd_work_group_size_i64(ptr addrsp
 
 ; Metadata uses i16 instead of i32
 ; CHECK-LABEL: @load_group_size_x_reqd_work_group_size_i16(
-; CHECK-NEXT: store i16 8,
+; CHECK: store i16 %group.size.x,
 define amdgpu_kernel void @load_group_size_x_reqd_work_group_size_i16(ptr addrspace(1) %out) #0 !reqd_work_group_size !3 {
   %dispatch.ptr = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr()
   %gep.group.size.x = getelementptr inbounds i8, ptr addrspace(4) %dispatch.ptr, i64 4
@@ -76,7 +76,7 @@ define amdgpu_kernel void @load_group_size_x_reqd_work_group_size_i16(ptr addrsp
 }
 
 ; CHECK-LABEL: @use_local_size_x_8_16_2(
-; CHECK-NEXT: store i64 8,
+; CHECK: store i64 %zext,
 define amdgpu_kernel void @use_local_size_x_8_16_2(ptr addrspace(1) %out) #0 !reqd_work_group_size !0 {
   %dispatch.ptr = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr()
   %gep.group.size.x = getelementptr inbounds i8, ptr addrspace(4) %dispatch.ptr, i64 4
@@ -94,7 +94,7 @@ define amdgpu_kernel void @use_local_size_x_8_16_2(ptr addrspace(1) %out) #0 !re
 }
 
 ; CHECK-LABEL: @use_local_size_y_8_16_2(
-; CHECK-NEXT: store i64 16,
+; CHECK: store i64 %zext,
 define amdgpu_kernel void @use_local_size_y_8_16_2(ptr addrspace(1) %out) #0 !reqd_work_group_size !0 {
   %dispatch.ptr = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr()
   %gep.group.size.y = getelementptr inbounds i8, ptr addrspace(4) %dispatch.ptr, i64 6
@@ -112,7 +112,7 @@ define amdgpu_kernel void @use_local_size_y_8_16_2(ptr addrspace(1) %out) #0 !re
 }
 
 ; CHECK-LABEL: @use_local_size_z_8_16_2(
-; CHECK-NEXT: store i64 2,
+; CHECK: store i64 %zext,
 define amdgpu_kernel void @use_local_size_z_8_16_2(ptr addrspace(1) %out) #0 !reqd_work_group_size !0 {
   %dispatch.ptr = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr()
   %gep.group.size.z = getelementptr inbounds i8, ptr addrspace(4) %dispatch.ptr, i64 8
@@ -134,7 +134,7 @@ define amdgpu_kernel void @use_local_size_z_8_16_2(ptr addrspace(1) %out) #0 !re
 
 ; CHECK-LABEL: @local_size_x_8_16_2_wrong_group_id(
 ; CHECK: %group.id = tail call i32 @llvm.amdgcn.workgroup.id.y()
-; CHECK: %group.id_x_group.size.x = shl i32 %group.id, 3
+; CHECK: %group.id_x_group.size.x = mul i32 %group.id, %group.size.x.zext
 define amdgpu_kernel void @local_size_x_8_16_2_wrong_group_id(ptr addrspace(1) %out) #0 !reqd_work_group_size !0 {
   %dispatch.ptr = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr()
   %gep.group.size.x = getelementptr inbounds i8, ptr addrspace(4) %dispatch.ptr, i64 4
@@ -154,7 +154,7 @@ define amdgpu_kernel void @local_size_x_8_16_2_wrong_group_id(ptr addrspace(1) %
 ; CHECK-LABEL: @local_size_x_8_16_2_wrong_grid_size(
 ; CHECK: %grid.size.x = load i32, ptr addrspace(4) %gep.grid.size.x, align 4
 ; CHECK: %group.id = tail call i32 @llvm.amdgcn.workgroup.id.x()
-; CHECK: %group.id_x_group.size.x = shl i32 %group.id, 3
+; CHECK: %group.id_x_group.size.x = mul i32 %group.id, %group.size.x.zext
   define amdgpu_kernel void @local_size_x_8_16_2_wrong_grid_size(ptr addrspace(1) %out) #0 !reqd_work_group_size !0 {
   %dispatch.ptr = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr()
   %gep.group.size.x = getelementptr inbounds i8, ptr addrspace(4) %dispatch.ptr, i64 4
@@ -174,9 +174,9 @@ define amdgpu_kernel void @local_size_x_8_16_2_wrong_group_id(ptr addrspace(1) %
 ; CHECK-LABEL: @local_size_x_8_16_2_wrong_cmp_type(
 ; CHECK: %grid.size.x = load i32, ptr addrspace(4) %gep.grid.size.x, align 4
 ; CHECK: %group.id = tail call i32 @llvm.amdgcn.workgroup.id.x()
-; CHECK: %group.id_x_group.size.x = shl i32 %group.id, 3
+; CHECK: %group.id_x_group.size.x = mul i32 %group.id, %group.size.x.zext
 ; CHECK: %sub = sub i32 %grid.size.x, %group.id_x_group.size.x
-; CHECK: %smin = call i32 @llvm.smin.i32(i32 %sub, i32 8)
+; CHECK: %smin = call i32 @llvm.smin.i32(i32 %sub, i32 %group.size.x.zext)
 define amdgpu_kernel void @local_size_x_8_16_2_wrong_cmp_type(ptr addrspace(1) %out) #0 !reqd_work_group_size !0 {
   %dispatch.ptr = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr()
   %gep.group.size.x = getelementptr inbounds i8, ptr addrspace(4) %dispatch.ptr, i64 4
@@ -194,9 +194,9 @@ define amdgpu_kernel void @local_size_x_8_16_2_wrong_cmp_type(ptr addrspace(1) %
 }
 
 ; CHECK-LABEL: @local_size_x_8_16_2_wrong_select(
-; CHECK: %group.id_x_group.size.x = shl i32 %group.id, 3
+; CHECK: %group.id_x_group.size.x = mul i32 %group.id, %group.size.x.zext
 ; CHECK: %sub = sub i32 %grid.size.x, %group.id_x_group.size.x
-; CHECK: %umax = call i32 @llvm.umax.i32(i32 %sub, i32 8)
+; CHECK: %umax = call i32 @llvm.umax.i32(i32 %sub, i32 %group.size.x.zext)
 ; CHECK: %zext = zext i32 %umax to i64
 define amdgpu_kernel void @local_size_x_8_16_2_wrong_select(ptr addrspace(1) %out) #0 !reqd_work_group_size !0 {
   %dispatch.ptr = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr()
@@ -218,7 +218,7 @@ define amdgpu_kernel void @local_size_x_8_16_2_wrong_select(ptr addrspace(1) %ou
 ; CHECK: %grid.size.x = load i16, ptr addrspace(4) %gep.grid.size.x, align 4
 ; CHECK: %grid.size.x.zext = zext i16 %grid.size.x to i32
 ; CHECK: %group.id = tail call i32 @llvm.amdgcn.workgroup.id.x()
-; CHECK: %group.id_x_group.size.x = shl i32 %group.id, 3
+; CHECK: %group.id_x_group.size.x = mul i32 %group.id, %group.size.x.zext
 ; CHECK: %sub = sub i32 %grid.size.x.zext, %group.id_x_group.size.x
 define amdgpu_kernel void @use_local_size_x_8_16_2_wrong_grid_load_size(ptr addrspace(1) %out) #0 !reqd_work_group_size !0 {
   %dispatch.ptr = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr()
@@ -238,7 +238,7 @@ define amdgpu_kernel void @use_local_size_x_8_16_2_wrong_grid_load_size(ptr addr
 }
 
 ; CHECK-LABEL: @func_group_size_x(
-; CHECK-NEXT: ret i32 8
+; CHECK: ret i32 %zext
 define i32 @func_group_size_x(ptr addrspace(1) %out) #0 !reqd_work_group_size !0 {
   %dispatch.ptr = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr()
   %gep.group.size.x = getelementptr inbounds i8, ptr addrspace(4) %dispatch.ptr, i64 4
@@ -248,7 +248,7 @@ define i32 @func_group_size_x(ptr addrspace(1) %out) #0 !reqd_work_group_size !0
 }
 
 ; CHECK-LABEL: @__ockl_get_local_size_reqd_size(
-; CHECK: %group.size = phi i32 [ 2, %bb17 ], [ 16, %bb9 ], [ 8, %bb1 ], [ 1, %bb ]
+; CHECK: %group.size = phi i16 [ %tmp24, %bb17 ], [ %tmp16, %bb9 ], [ %tmp8, %bb1 ], [ 1, %bb ]
 define i64 @__ockl_get_local_size_reqd_size(i32 %arg) #1 !reqd_work_group_size !0 {
 bb:
   %tmp = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr() #2
@@ -295,9 +295,9 @@ bb25:                                             ; preds = %bb17, %bb9, %bb1, %
 }
 
 ; CHECK-LABEL: @all_local_size(
-; CHECK-NEXT: store volatile i64 8, ptr addrspace(1) %out, align 4
-; CHECK-NEXT: store volatile i64 16, ptr addrspace(1) %out, align 4
-; CHECK-NEXT: store volatile i64 2, ptr addrspace(1) %out, align 4
+; CHECK: store volatile i64 %tmp34.i, ptr addrspace(1) %out, align 4
+; CHECK-NEXT: store volatile i64 %tmp34.i14, ptr addrspace(1) %out, align 4
+; CHECK-NEXT: store volatile i64 %tmp34.i7, ptr addrspace(1) %out, align 4
 define amdgpu_kernel void @all_local_size(ptr addrspace(1) nocapture readnone %out) #0 !reqd_work_group_size !0 {
   %tmp.i = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr() #0
   %tmp2.i = tail call i32 @llvm.amdgcn.workgroup.id.x() #0
@@ -376,8 +376,8 @@ define amdgpu_kernel void @load_group_size_xy_i32(ptr addrspace(1) %out) #0 !req
 }
 
 ; CHECK-LABEL: @load_group_size_x_y_multiple_dispatch_ptr(
-; CHECK-NEXT: store volatile i16 8, ptr addrspace(1) %out, align 2
-; CHECK-NEXT: store volatile i16 16, ptr addrspace(1) %out, align 2
+; CHECK: store volatile i16 %group.size.x, ptr addrspace(1) %out, align 2
+; CHECK: store volatile i16 %group.size.y, ptr addrspace(1) %out, align 2
 define amdgpu_kernel void @load_group_size_x_y_multiple_dispatch_ptr(ptr addrspace(1) %out) #0 !reqd_work_group_size !0 {
   %dispatch.ptr0 = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr()
   %gep.group.size.x = getelementptr inbounds i8, ptr addrspace(4) %dispatch.ptr0, i64 4
@@ -396,8 +396,8 @@ define amdgpu_kernel void @load_group_size_x_y_multiple_dispatch_ptr(ptr addrspa
 ; CHECK-NEXT: %dispatch.ptr = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr()
 ; CHECK-NEXT: %gep.group.size.x = getelementptr inbounds i8, ptr addrspace(4) %dispatch.ptr, i64 4
 ; CHECK-NEXT: %group.size.x = load i16, ptr addrspace(4) %gep.group.size.x, align 4
-; CHECK-NEXT: %zext = zext i16 %group.size.x to i64
-; CHECK-NEXT: store i64 %zext, ptr addrspace(1) %out, align 4
+; CHECK: %group.size.x.zext = zext i16 %group.size.x to i32
+; CHECK: store i64 %zext, ptr addrspace(1) %out
 define amdgpu_kernel void @use_local_size_x_uniform_work_group_size(ptr addrspace(1) %out) #2 {
   %dispatch.ptr = tail call ptr addrspace(4) @llvm.amdgcn.dispatch.ptr()
   %gep.group.size.x = getelementptr inbounds i8, ptr addrspace(4) %dispatch.ptr, i64 4
@@ -456,3 +456,6 @@ attributes #3 = { nounwind "uniform-work-group-size"="false" }
 !1 = !{i32 8, i32 16}
 !2 = !{i64 8, i64 16, i64 2}
 !3 = !{i16 8, i16 16, i16 2}
+
+!llvm.module.flags = !{!4}
+!4 = !{i32 1, !"amdgpu_code_object_version", i32 500}
diff --git a/llvm/test/CodeGen/AMDGPU/simple-indirect-call.ll b/llvm/test/CodeGen/AMDGPU/simple-indirect-call.ll
index dcc90c0dcd40..e7c5aaf043ef 100644
--- a/llvm/test/CodeGen/AMDGPU/simple-indirect-call.ll
+++ b/llvm/test/CodeGen/AMDGPU/simple-indirect-call.ll
@@ -43,9 +43,9 @@ define amdgpu_kernel void @test_simple_indirect_call() {
 ; GFX9-LABEL: test_simple_indirect_call:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x4
-; GFX9-NEXT:    s_add_u32 flat_scratch_lo, s12, s17
-; GFX9-NEXT:    s_addc_u32 flat_scratch_hi, s13, 0
-; GFX9-NEXT:    s_add_u32 s0, s0, s17
+; GFX9-NEXT:    s_add_u32 flat_scratch_lo, s10, s15
+; GFX9-NEXT:    s_addc_u32 flat_scratch_hi, s11, 0
+; GFX9-NEXT:    s_add_u32 s0, s0, s15
 ; GFX9-NEXT:    s_addc_u32 s1, s1, 0
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    s_lshr_b32 s4, s4, 16
@@ -76,3 +76,6 @@ define amdgpu_kernel void @test_simple_indirect_call() {
 ; ATTRIBUTOR_GCN: attributes #[[ATTR0]] = { "amdgpu-no-completion-action" "amdgpu-no-default-queue" "amdgpu-no-dispatch-id" "amdgpu-no-dispatch-ptr" "amdgpu-no-heap-ptr" "amdgpu-no-hostcall-ptr" "amdgpu-no-implicitarg-ptr" "amdgpu-no-lds-kernel-id" "amdgpu-no-multigrid-sync-arg" "amdgpu-no-queue-ptr" "amdgpu-no-workgroup-id-x" "amdgpu-no-workgroup-id-y" "amdgpu-no-workgroup-id-z" "amdgpu-no-workitem-id-x" "amdgpu-no-workitem-id-y" "amdgpu-no-workitem-id-z" "amdgpu-waves-per-eu"="4,10" "uniform-work-group-size"="false" }
 ; ATTRIBUTOR_GCN: attributes #[[ATTR1]] = { "uniform-work-group-size"="false" }
 ;.
+
+!llvm.module.flags = !{!0}
+!0 = !{i32 1, !"amdgpu_code_object_version", i32 500}