Skip to content

AMDGPU: Replace test uses of ptr addrspace(5) undef with poison #131101

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Mar 14, 2025
Merged

AMDGPU: Replace test uses of ptr addrspace(5) undef with poison #131101

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion llvm/test/CodeGen/AMDGPU/GlobalISel/divergent-control-flow.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -183,7 +183,7 @@ bb8:
br i1 %tmp10, label %bb11, label %bb12

bb11:
store float 4.0, ptr addrspace(5) undef, align 4
store float 4.0, ptr addrspace(5) poison, align 4
br label %bb12

bb12:
Expand Down
4 changes: 2 additions & 2 deletions llvm/test/CodeGen/AMDGPU/collapse-endcf.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1360,7 +1360,7 @@ bb2: ; preds = %bb1
br i1 %tmp3, label %bb4, label %bb10

bb4: ; preds = %bb2
%tmp6 = load float, ptr addrspace(5) undef
%tmp6 = load float, ptr addrspace(5) poison
%tmp7 = fcmp olt float %tmp6, 0.0
br i1 %tmp7, label %bb8, label %Flow

Expand All @@ -1380,7 +1380,7 @@ Flow1: ; preds = %bb10
br label %bb1

bb12: ; preds = %bb10
store volatile <4 x float> %tmp11, ptr addrspace(5) undef, align 16
store volatile <4 x float> %tmp11, ptr addrspace(5) poison, align 16
ret void
}

Expand Down
2 changes: 1 addition & 1 deletion llvm/test/CodeGen/AMDGPU/dag-divergence.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@
; GCN: flat_load_dword
; GCN-NOT: s_load_dword s
define amdgpu_kernel void @private_load_maybe_divergent(ptr addrspace(4) %k, ptr %flat) {
%load = load volatile i32, ptr addrspace(5) undef, align 4
%load = load volatile i32, ptr addrspace(5) poison, align 4
%gep = getelementptr inbounds i32, ptr addrspace(4) %k, i32 %load
%maybe.not.uniform.load = load i32, ptr addrspace(4) %gep, align 4
store i32 %maybe.not.uniform.load, ptr addrspace(1) poison
Expand Down
2 changes: 1 addition & 1 deletion llvm/test/CodeGen/AMDGPU/debug-value.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -35,7 +35,7 @@ bb25: ; preds = %bb

bb28: ; preds = %bb25, %bb21
%tmp29 = phi <4 x float> [ %tmp27, %bb25 ], [ %tmp24, %bb21 ]
store <4 x float> %tmp29, ptr addrspace(5) undef, align 16
store <4 x float> %tmp29, ptr addrspace(5) poison, align 16
%tmp30 = getelementptr inbounds %struct.wombat, ptr addrspace(1) %arg, i64 %tmp2, i32 2, i64 2
%tmp31 = load i32, ptr addrspace(1) %tmp30, align 4
%tmp32 = sext i32 %tmp31 to i64
Expand Down
2 changes: 1 addition & 1 deletion llvm/test/CodeGen/AMDGPU/frame-index-elimination.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -255,7 +255,7 @@ define void @undefined_stack_store_reg(float %arg, i32 %arg1) #0 {
bb:
%tmp = alloca <4 x float>, align 16, addrspace(5)
%tmp2 = insertelement <4 x float> poison, float %arg, i32 0
store <4 x float> %tmp2, ptr addrspace(5) undef
store <4 x float> %tmp2, ptr addrspace(5) poison
%tmp3 = icmp eq i32 %arg1, 0
br i1 %tmp3, label %bb4, label %bb5

Expand Down
2 changes: 1 addition & 1 deletion llvm/test/CodeGen/AMDGPU/lds-dma-waitcnt.mir
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -77,7 +77,7 @@ name: scratch_load_lds_dword_ds_read
body: |
bb.0:
$m0 = S_MOV_B32 0
SCRATCH_LOAD_LDS_DWORD $vgpr0, 4, 0, implicit $exec, implicit $m0 :: (load (s32) from `ptr addrspace(5) undef` + 4), (store (s32) into `ptr addrspace(3) undef` + 4)
SCRATCH_LOAD_LDS_DWORD $vgpr0, 4, 0, implicit $exec, implicit $m0 :: (load (s32) from `ptr addrspace(5) poison` + 4), (store (s32) into `ptr addrspace(3) undef` + 4)
$vgpr0 = DS_READ_B32_gfx9 $vgpr1, 0, 0, implicit $m0, implicit $exec :: (load (s32) from `ptr addrspace(3) undef`)
S_ENDPGM 0
Expand Down
2 changes: 1 addition & 1 deletion llvm/test/CodeGen/AMDGPU/mdt-preserving-crash.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -104,7 +104,7 @@ if.end5.i362: ; preds = %if.then3.i356, %if.
%conv612.i359 = sext i8 %5 to i32
%sub13.i360 = add nsw i32 %conv612.i359, -48
%cmp714.i361 = icmp ugt i32 %sub13.i360, 9
store i8 0, ptr addrspace(5) undef, align 16
store i8 0, ptr addrspace(5) poison, align 16
%6 = load i8, ptr addrspace(1) getelementptr inbounds ([4096 x i8], ptr addrspace(1) @_RSENC_gDcd_______________________________, i64 0, i64 1153), align 1
%arrayidx232250.1 = getelementptr inbounds [128 x i8], ptr addrspace(5) %pD10, i32 0, i32 1
store i8 %6, ptr addrspace(5) %arrayidx232250.1, align 1
Expand Down
6 changes: 3 additions & 3 deletions llvm/test/CodeGen/AMDGPU/memory-legalizer-multiple-mem-operands-atomics.mir
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,13 +23,13 @@ body: |
$sgpr10 = S_MOV_B32 4294967295, implicit-def $sgpr8_sgpr9_sgpr10_sgpr11
$sgpr11 = S_MOV_B32 15204352, implicit-def $sgpr8_sgpr9_sgpr10_sgpr11
$vgpr0 = V_MOV_B32_e32 1, implicit $exec
BUFFER_STORE_DWORD_OFFSET killed $vgpr0, $sgpr8_sgpr9_sgpr10_sgpr11, $sgpr3, 4, 0, 0, implicit $exec :: (store (s32) into `ptr addrspace(5) undef`)
BUFFER_STORE_DWORD_OFFSET killed $vgpr0, $sgpr8_sgpr9_sgpr10_sgpr11, $sgpr3, 4, 0, 0, implicit $exec :: (store (s32) into `ptr addrspace(5) poison`)
S_WAITCNT 127
S_CMP_LG_U32 killed $sgpr2, 0, implicit-def $scc
S_WAITCNT 3855
$vgpr0 = V_MOV_B32_e32 2, implicit $exec
$vgpr1 = V_MOV_B32_e32 32772, implicit $exec
BUFFER_STORE_DWORD_OFFEN killed $vgpr0, killed $vgpr1, $sgpr8_sgpr9_sgpr10_sgpr11, $sgpr3, 0, 0, 0, implicit $exec :: (store (s32) into `ptr addrspace(5) undef`)
BUFFER_STORE_DWORD_OFFEN killed $vgpr0, killed $vgpr1, $sgpr8_sgpr9_sgpr10_sgpr11, $sgpr3, 0, 0, 0, implicit $exec :: (store (s32) into `ptr addrspace(5) poison`)
S_CBRANCH_SCC0 %bb.1, implicit killed $scc

bb.2:
Expand All @@ -55,7 +55,7 @@ body: |
S_WAITCNT 127
$sgpr0 = S_LSHL_B32 killed $sgpr0, 2, implicit-def dead $scc
$vgpr0 = V_ADD_CO_U32_e32 killed $sgpr0, killed $vgpr0, implicit-def dead $vcc, implicit $exec
$vgpr0 = BUFFER_LOAD_DWORD_OFFEN killed $vgpr0, killed $sgpr8_sgpr9_sgpr10_sgpr11, $sgpr3, 0, 0, 0, implicit $exec :: (load syncscope("agent-one-as") unordered (s32) from `ptr addrspace(1) undef`), (load syncscope("workgroup-one-as") seq_cst (s32) from `ptr addrspace(5) undef`)
$vgpr0 = BUFFER_LOAD_DWORD_OFFEN killed $vgpr0, killed $sgpr8_sgpr9_sgpr10_sgpr11, $sgpr3, 0, 0, 0, implicit $exec :: (load syncscope("agent-one-as") unordered (s32) from `ptr addrspace(1) undef`), (load syncscope("workgroup-one-as") seq_cst (s32) from `ptr addrspace(5) poison`)
$vgpr1 = V_MOV_B32_e32 $sgpr4, implicit $exec, implicit-def $vgpr1_vgpr2, implicit $sgpr4_sgpr5
$vgpr2 = V_MOV_B32_e32 killed $sgpr5, implicit $exec, implicit $sgpr4_sgpr5, implicit $exec
S_WAITCNT 3952
Expand Down
2 changes: 1 addition & 1 deletion llvm/test/CodeGen/AMDGPU/move-to-valu-worklist.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@
; GCN-NEXT: v_and_b32_e32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}}
define amdgpu_kernel void @in_worklist_once() #0 {
bb:
%tmp = load i64, ptr addrspace(5) undef
%tmp = load i64, ptr addrspace(5) poison
br label %bb1

bb1: ; preds = %bb1, %bb
Expand Down
2 changes: 1 addition & 1 deletion llvm/test/CodeGen/AMDGPU/operand-folding.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -130,7 +130,7 @@ define amdgpu_kernel void @no_fold_tied_subregister() #1 {
; CHECK: v_xor_b32_e32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}}
define void @no_extra_fold_on_same_opnd() #1 {
entry:
%s0 = load i32, ptr addrspace(5) undef, align 4
%s0 = load i32, ptr addrspace(5) poison, align 4
%s0.i64= zext i32 %s0 to i64
br label %for.body.i.i

Expand Down
4 changes: 2 additions & 2 deletions llvm/test/CodeGen/AMDGPU/private-access-no-objects.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,7 @@
; OPTNONE-NOT: s_mov_b32
; OPTNONE: buffer_store_dword v{{[0-9]+}}, v{{[0-9]+}}, s[0:3], 0 offen{{$}}
define amdgpu_kernel void @store_to_undef() #0 {
store volatile i32 0, ptr addrspace(5) undef
store volatile i32 0, ptr addrspace(5) poison
ret void
}

Expand All @@ -36,7 +36,7 @@ define amdgpu_kernel void @store_to_inttoptr() #0 {
; OPT-DAG: s_mov_b64 s{{\[[0-9]+}}:[[RSRC_HI:[0-9]+]]], s[2:3]
; OPT: buffer_load_dword v{{[0-9]+}}, v{{[0-9]+}}, s[[[RSRC_LO]]:[[RSRC_HI]]], 0 offen glc{{$}}
define amdgpu_kernel void @load_from_undef() #0 {
%ld = load volatile i32, ptr addrspace(5) undef
%ld = load volatile i32, ptr addrspace(5) poison
ret void
}

Expand Down
6 changes: 3 additions & 3 deletions llvm/test/CodeGen/AMDGPU/promote-alloca-to-lds-select.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,10 +4,10 @@ target datalayout = "e-p:64:64-p1:64:64-p2:32:32-p3:32:32-p4:64:64-p5:32:32-p6:3

; CHECK-LABEL: @lds_promoted_alloca_select_invalid_pointer_operand(
; CHECK: %alloca = alloca i32
; CHECK: select i1 undef, ptr addrspace(5) undef, ptr addrspace(5) %alloca
; CHECK: select i1 undef, ptr addrspace(5) poison, ptr addrspace(5) %alloca
define amdgpu_kernel void @lds_promoted_alloca_select_invalid_pointer_operand() #0 {
%alloca = alloca i32, align 4, addrspace(5)
%select = select i1 undef, ptr addrspace(5) undef, ptr addrspace(5) %alloca
%select = select i1 undef, ptr addrspace(5) poison, ptr addrspace(5) %alloca
store i32 0, ptr addrspace(5) %select, align 4
ret void
}
Expand Down Expand Up @@ -87,7 +87,7 @@ entry:

bb1:
%ptr2 = getelementptr inbounds [16 x i32], ptr addrspace(5) %alloca, i32 0, i32 %c
%select0 = select i1 undef, ptr addrspace(5) undef, ptr addrspace(5) %ptr2
%select0 = select i1 undef, ptr addrspace(5) poison, ptr addrspace(5) %ptr2
store i32 0, ptr addrspace(5) %ptr1
br label %bb2

Expand Down
12 changes: 6 additions & 6 deletions llvm/test/CodeGen/AMDGPU/sad.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -105,7 +105,7 @@ define amdgpu_kernel void @v_sad_u32_multi_use_sub_pat1(ptr addrspace(1) %out, i
%t1 = select i1 %icmp1, i32 %a, i32 %b

%ret0 = sub i32 %t0, %t1
store volatile i32 %ret0, ptr addrspace(5) undef
store volatile i32 %ret0, ptr addrspace(5) poison
%ret = add i32 %ret0, %c

store i32 %ret, ptr addrspace(1) %out
Expand Down Expand Up @@ -139,7 +139,7 @@ define amdgpu_kernel void @v_sad_u32_multi_use_add_pat1(ptr addrspace(1) %out, i

%ret0 = sub i32 %t0, %t1
%ret = add i32 %ret0, %c
store volatile i32 %ret, ptr addrspace(5) undef
store volatile i32 %ret, ptr addrspace(5) poison
store i32 %ret, ptr addrspace(1) %out
ret void
}
Expand Down Expand Up @@ -167,7 +167,7 @@ define amdgpu_kernel void @v_sad_u32_multi_use_max_pat1(ptr addrspace(1) %out, i
; GCN-NEXT: s_endpgm
%icmp0 = icmp ugt i32 %a, %b
%t0 = select i1 %icmp0, i32 %a, i32 %b
store volatile i32 %t0, ptr addrspace(5) undef
store volatile i32 %t0, ptr addrspace(5) poison

%icmp1 = icmp ule i32 %a, %b
%t1 = select i1 %icmp1, i32 %a, i32 %b
Expand Down Expand Up @@ -206,7 +206,7 @@ define amdgpu_kernel void @v_sad_u32_multi_use_min_pat1(ptr addrspace(1) %out, i
%icmp1 = icmp ule i32 %a, %b
%t1 = select i1 %icmp1, i32 %a, i32 %b

store volatile i32 %t1, ptr addrspace(5) undef
store volatile i32 %t1, ptr addrspace(5) poison

%ret0 = sub i32 %t0, %t1
%ret = add i32 %ret0, %c
Expand Down Expand Up @@ -238,7 +238,7 @@ define amdgpu_kernel void @v_sad_u32_multi_use_sub_pat2(ptr addrspace(1) %out, i
; GCN-NEXT: s_endpgm
%icmp0 = icmp ugt i32 %a, %b
%sub0 = sub i32 %a, %b
store volatile i32 %sub0, ptr addrspace(5) undef
store volatile i32 %sub0, ptr addrspace(5) poison
%sub1 = sub i32 %b, %a
%ret0 = select i1 %icmp0, i32 %sub0, i32 %sub1

Expand Down Expand Up @@ -274,7 +274,7 @@ define amdgpu_kernel void @v_sad_u32_multi_use_select_pat2(ptr addrspace(1) %out
%sub0 = sub i32 %a, %b
%sub1 = sub i32 %b, %a
%ret0 = select i1 %icmp0, i32 %sub0, i32 %sub1
store volatile i32 %ret0, ptr addrspace(5) undef
store volatile i32 %ret0, ptr addrspace(5) poison

%ret = add i32 %ret0, %c

Expand Down
8 changes: 4 additions & 4 deletions llvm/test/CodeGen/AMDGPU/scalar_to_vector_v2x16.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,18 +5,18 @@
; GCN: v_mov_b32_e32 [[V:v[0-9]+]], 42
; GCN: buffer_store_short [[V]],
define void @scalar_to_vector_i16() {
%tmp = load <2 x i16>, ptr addrspace(5) undef
%tmp = load <2 x i16>, ptr addrspace(5) poison
%tmp1 = insertelement <2 x i16> %tmp, i16 42, i64 0
store <2 x i16> %tmp1, ptr addrspace(5) undef
store <2 x i16> %tmp1, ptr addrspace(5) poison
ret void
}

; GCN-LABEL: {{^}}scalar_to_vector_f16:
; GCN: v_mov_b32_e32 [[V:v[0-9]+]], 0x3c00
; GCN: buffer_store_short [[V]],
define void @scalar_to_vector_f16() {
%tmp = load <2 x half>, ptr addrspace(5) undef
%tmp = load <2 x half>, ptr addrspace(5) poison
%tmp1 = insertelement <2 x half> %tmp, half 1.0, i64 0
store <2 x half> %tmp1, ptr addrspace(5) undef
store <2 x half> %tmp1, ptr addrspace(5) poison
ret void
}
2 changes: 1 addition & 1 deletion llvm/test/CodeGen/AMDGPU/stack-pointer-offset-relative-frameindex.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -148,7 +148,7 @@ if.then4.i: ; preds = %entry
%add1.i.i = add i32 %add.i.i, 0
%mul.i.i.i.i = mul i32 %add1.i.i, 1103515245
%add.i.i.i.i = add i32 %mul.i.i.i.i, 12345
store i32 %add.i.i.i.i, ptr addrspace(5) undef, align 16
store i32 %add.i.i.i.i, ptr addrspace(5) poison, align 16
br label %shader_eval_surface.exit

shader_eval_surface.exit: ; preds = %entry
Expand Down
Loading