https://github.com/arsenm updated
https://github.com/llvm/llvm-project/pull/130904
>From fa3c82be14f0e94ea7e1a33c167968c7379f2563 Mon Sep 17 00:00:00 2001
From: Matt Arsenault <matthew.arsena...@amd.com>
Date: Wed, 12 Mar 2025 13:24:50 +0700
Subject: [PATCH] AMDGPU: Replace ptr addrspace(8) undef uses with poison
---
llvm/test/CodeGen/AMDGPU/amdpal.ll | 2 +-
.../CodeGen/AMDGPU/combine-add-zext-xor.ll | 12 +-
llvm/test/CodeGen/AMDGPU/else.ll | 2 +-
.../AMDGPU/extract_subvector_vec4_vec3.ll | 8 +-
llvm/test/CodeGen/AMDGPU/llvm.amdgcn.exp.ll | 2 +-
.../llvm.amdgcn.raw.ptr.buffer.atomic.ll | 2 +-
.../llvm.amdgcn.struct.ptr.buffer.atomic.ll | 2 +-
.../test/CodeGen/AMDGPU/loop_exit_with_xor.ll | 6 +-
.../lower-work-group-id-intrinsics-hsa.ll | 2 +-
.../lower-work-group-id-intrinsics-pal.ll | 2 +-
llvm/test/CodeGen/AMDGPU/merge-store-crash.ll | 2 +-
.../test/CodeGen/AMDGPU/merge-store-usedef.ll | 2 +-
.../AMDGPU/required-export-priority.ll | 2 +-
.../AMDGPU/si-triv-disjoint-mem-access.ll | 2 +-
llvm/test/CodeGen/AMDGPU/wave32.ll | 8 +-
llvm/test/CodeGen/AMDGPU/wqm.ll | 110 +++++++++---------
16 files changed, 83 insertions(+), 83 deletions(-)
diff --git a/llvm/test/CodeGen/AMDGPU/amdpal.ll
b/llvm/test/CodeGen/AMDGPU/amdpal.ll
index 171df029615ed..fd9227d2f4319 100644
--- a/llvm/test/CodeGen/AMDGPU/amdpal.ll
+++ b/llvm/test/CodeGen/AMDGPU/amdpal.ll
@@ -72,7 +72,7 @@ entry:
%e = getelementptr [2 x i32], ptr addrspace(5) %v1, i32 0, i32 %idx
%x = load i32, ptr addrspace(5) %e
%xf = bitcast i32 %x to float
- call void @llvm.amdgcn.raw.ptr.buffer.store.f32(float %xf, ptr addrspace(8)
undef, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.raw.ptr.buffer.store.f32(float %xf, ptr addrspace(8)
poison, i32 0, i32 0, i32 0)
ret void
}
diff --git a/llvm/test/CodeGen/AMDGPU/combine-add-zext-xor.ll
b/llvm/test/CodeGen/AMDGPU/combine-add-zext-xor.ll
index b42542db6dbd8..f8227f0039af7 100644
--- a/llvm/test/CodeGen/AMDGPU/combine-add-zext-xor.ll
+++ b/llvm/test/CodeGen/AMDGPU/combine-add-zext-xor.ll
@@ -66,7 +66,7 @@ define i32 @combine_add_zext_xor() {
br i1 undef, label %bb9, label %bb
bb: ; preds = %.a
- %.i3 = call i32 @llvm.amdgcn.raw.ptr.buffer.load.i32(ptr addrspace(8) undef,
i32 %.2, i32 64, i32 1)
+ %.i3 = call i32 @llvm.amdgcn.raw.ptr.buffer.load.i32(ptr addrspace(8)
poison, i32 %.2, i32 64, i32 1)
%i5 = icmp eq i32 %.i3, 0
br label %bb9
@@ -146,7 +146,7 @@ define i32 @combine_sub_zext_xor() {
br i1 undef, label %bb9, label %bb
bb: ; preds = %.a
- %.i3 = call i32 @llvm.amdgcn.raw.ptr.buffer.load.i32(ptr addrspace(8) undef,
i32 %.2, i32 64, i32 1)
+ %.i3 = call i32 @llvm.amdgcn.raw.ptr.buffer.load.i32(ptr addrspace(8)
poison, i32 %.2, i32 64, i32 1)
%i5 = icmp eq i32 %.i3, 0
br label %bb9
@@ -229,7 +229,7 @@ define i32 @combine_add_zext_or() {
br i1 undef, label %bb9, label %bb
bb: ; preds = %.a
- %.i3 = call i32 @llvm.amdgcn.raw.ptr.buffer.load.i32(ptr addrspace(8) undef,
i32 %.2, i32 64, i32 1)
+ %.i3 = call i32 @llvm.amdgcn.raw.ptr.buffer.load.i32(ptr addrspace(8)
poison, i32 %.2, i32 64, i32 1)
%i5 = icmp eq i32 %.i3, 0
br label %bb9
@@ -313,7 +313,7 @@ define i32 @combine_sub_zext_or() {
br i1 undef, label %bb9, label %bb
bb: ; preds = %.a
- %.i3 = call i32 @llvm.amdgcn.raw.ptr.buffer.load.i32(ptr addrspace(8) undef,
i32 %.2, i32 64, i32 1)
+ %.i3 = call i32 @llvm.amdgcn.raw.ptr.buffer.load.i32(ptr addrspace(8)
poison, i32 %.2, i32 64, i32 1)
%i5 = icmp eq i32 %.i3, 0
br label %bb9
@@ -392,7 +392,7 @@ define i32 @combine_add_zext_and() {
br i1 undef, label %bb9, label %bb
bb: ; preds = %.a
- %.i3 = call i32 @llvm.amdgcn.raw.ptr.buffer.load.i32(ptr addrspace(8) undef,
i32 %.2, i32 64, i32 1)
+ %.i3 = call i32 @llvm.amdgcn.raw.ptr.buffer.load.i32(ptr addrspace(8)
poison, i32 %.2, i32 64, i32 1)
%i5 = icmp eq i32 %.i3, 0
br label %bb9
@@ -471,7 +471,7 @@ define i32 @combine_sub_zext_and() {
br i1 undef, label %bb9, label %bb
bb: ; preds = %.a
- %.i3 = call i32 @llvm.amdgcn.raw.ptr.buffer.load.i32(ptr addrspace(8) undef,
i32 %.2, i32 64, i32 1)
+ %.i3 = call i32 @llvm.amdgcn.raw.ptr.buffer.load.i32(ptr addrspace(8)
poison, i32 %.2, i32 64, i32 1)
%i5 = icmp eq i32 %.i3, 0
br label %bb9
diff --git a/llvm/test/CodeGen/AMDGPU/else.ll b/llvm/test/CodeGen/AMDGPU/else.ll
index 4a3018e67b17d..884f5305407a1 100644
--- a/llvm/test/CodeGen/AMDGPU/else.ll
+++ b/llvm/test/CodeGen/AMDGPU/else.ll
@@ -47,7 +47,7 @@ else:
end:
%r = phi float [ %v.if, %if ], [ %v.else, %else ]
- call void @llvm.amdgcn.raw.ptr.buffer.store.f32(float %r, ptr addrspace(8)
undef, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.raw.ptr.buffer.store.f32(float %r, ptr addrspace(8)
poison, i32 0, i32 0, i32 0)
ret void
}
diff --git a/llvm/test/CodeGen/AMDGPU/extract_subvector_vec4_vec3.ll
b/llvm/test/CodeGen/AMDGPU/extract_subvector_vec4_vec3.ll
index a44231f39ae6f..37a2e9e24dbd6 100644
--- a/llvm/test/CodeGen/AMDGPU/extract_subvector_vec4_vec3.ll
+++ b/llvm/test/CodeGen/AMDGPU/extract_subvector_vec4_vec3.ll
@@ -13,7 +13,7 @@ define amdgpu_hs void @main(ptr addrspace(6) inreg %arg) {
; GCN-NEXT: [[DEF:%[0-9]+]]:sreg_32 = IMPLICIT_DEF
; GCN-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[DEF]]
; GCN-NEXT: [[DEF1:%[0-9]+]]:sgpr_128 = IMPLICIT_DEF
- ; GCN-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 =
BUFFER_LOAD_DWORDX4_OFFEN [[COPY]], [[DEF1]], [[S_MOV_B32_]], 0, 0, 0, implicit
$exec :: (dereferenceable load (s128) from `ptr addrspace(8) undef`, align 1,
addrspace 8)
+ ; GCN-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 =
BUFFER_LOAD_DWORDX4_OFFEN [[COPY]], [[DEF1]], [[S_MOV_B32_]], 0, 0, 0, implicit
$exec :: (dereferenceable load (s128) from `ptr addrspace(8) poison`, align 1,
addrspace 8)
; GCN-NEXT: [[COPY1:%[0-9]+]]:sgpr_32 = COPY
[[BUFFER_LOAD_DWORDX4_OFFEN]].sub2
; GCN-NEXT: [[COPY2:%[0-9]+]]:sgpr_32 = COPY
[[BUFFER_LOAD_DWORDX4_OFFEN]].sub1
; GCN-NEXT: [[COPY3:%[0-9]+]]:sgpr_32 = COPY
[[BUFFER_LOAD_DWORDX4_OFFEN]].sub0
@@ -22,14 +22,14 @@ define amdgpu_hs void @main(ptr addrspace(6) inreg %arg) {
; GCN-NEXT: [[DEF2:%[0-9]+]]:sreg_32 = IMPLICIT_DEF
; GCN-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[DEF2]]
; GCN-NEXT: [[DEF3:%[0-9]+]]:sgpr_128 = IMPLICIT_DEF
- ; GCN-NEXT: BUFFER_STORE_DWORDX3_OFFEN_exact killed [[COPY4]], [[COPY5]],
[[DEF3]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable store
(s96) into `ptr addrspace(8) undef`, align 1, addrspace 8)
+ ; GCN-NEXT: BUFFER_STORE_DWORDX3_OFFEN_exact killed [[COPY4]], [[COPY5]],
[[DEF3]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable store
(s96) into `ptr addrspace(8) poison`, align 1, addrspace 8)
; GCN-NEXT: S_ENDPGM 0
main_body:
- %tmp25 = call <4 x float> @llvm.amdgcn.raw.ptr.buffer.load.v4f32(ptr
addrspace(8) undef, i32 undef, i32 0, i32 0)
+ %tmp25 = call <4 x float> @llvm.amdgcn.raw.ptr.buffer.load.v4f32(ptr
addrspace(8) poison, i32 undef, i32 0, i32 0)
%tmp27 = bitcast <4 x float> %tmp25 to <16 x i8>
%tmp28 = shufflevector <16 x i8> %tmp27, <16 x i8> poison, <12 x i32> <i32
0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32
11>
%tmp29 = bitcast <12 x i8> %tmp28 to <3 x i32>
- call void @llvm.amdgcn.raw.ptr.buffer.store.v3i32(<3 x i32> %tmp29, ptr
addrspace(8) undef, i32 undef, i32 0, i32 0) #3
+ call void @llvm.amdgcn.raw.ptr.buffer.store.v3i32(<3 x i32> %tmp29, ptr
addrspace(8) poison, i32 undef, i32 0, i32 0) #3
ret void
}
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.exp.ll
b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.exp.ll
index c120c58ce6f33..c506e08855470 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.exp.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.exp.ll
@@ -623,7 +623,7 @@ define amdgpu_kernel void
@test_export_pos_before_param_ordered(float %x, float
define amdgpu_kernel void @test_export_pos_before_param_across_load(i32 %idx)
#0 {
call void @llvm.amdgcn.exp.f32(i32 32, i32 15, float 1.0, float 1.0, float
1.0, float 1.0, i1 false, i1 false)
call void @llvm.amdgcn.exp.f32(i32 33, i32 15, float 1.0, float 1.0, float
1.0, float 0.5, i1 false, i1 false)
- %load = call float @llvm.amdgcn.raw.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx, i32 0, i32 0)
+ %load = call float @llvm.amdgcn.raw.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx, i32 0, i32 0)
call void @llvm.amdgcn.exp.f32(i32 12, i32 15, float 0.0, float 0.0, float
0.0, float %load, i1 true, i1 false)
ret void
}
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.raw.ptr.buffer.atomic.ll
b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.raw.ptr.buffer.atomic.ll
index 9c9cf36baecf1..843ad563df69e 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.raw.ptr.buffer.atomic.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.raw.ptr.buffer.atomic.ll
@@ -103,7 +103,7 @@ main_body:
;CHECK: buffer_atomic_add v0,
define amdgpu_ps float @test4() {
main_body:
- %v = call i32 @llvm.amdgcn.raw.ptr.buffer.atomic.add.i32(i32 1, ptr
addrspace(8) undef, i32 4, i32 0, i32 0)
+ %v = call i32 @llvm.amdgcn.raw.ptr.buffer.atomic.add.i32(i32 1, ptr
addrspace(8) poison, i32 4, i32 0, i32 0)
%v.float = bitcast i32 %v to float
ret float %v.float
}
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.struct.ptr.buffer.atomic.ll
b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.struct.ptr.buffer.atomic.ll
index fb44515595a55..607f6001f99b4 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.struct.ptr.buffer.atomic.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.struct.ptr.buffer.atomic.ll
@@ -115,7 +115,7 @@ main_body:
;CHECK: buffer_atomic_add v0,
define amdgpu_ps float @test4() {
main_body:
- %v = call i32 @llvm.amdgcn.struct.ptr.buffer.atomic.add.i32(i32 1, ptr
addrspace(8) undef, i32 0, i32 4, i32 0, i32 0)
+ %v = call i32 @llvm.amdgcn.struct.ptr.buffer.atomic.add.i32(i32 1, ptr
addrspace(8) poison, i32 0, i32 4, i32 0, i32 0)
%v.float = bitcast i32 %v to float
ret float %v.float
}
diff --git a/llvm/test/CodeGen/AMDGPU/loop_exit_with_xor.ll
b/llvm/test/CodeGen/AMDGPU/loop_exit_with_xor.ll
index a407cd20bf762..2ce612b176587 100644
--- a/llvm/test/CodeGen/AMDGPU/loop_exit_with_xor.ll
+++ b/llvm/test/CodeGen/AMDGPU/loop_exit_with_xor.ll
@@ -46,7 +46,7 @@ loop:
br i1 %tmp27, label %then, label %endif
then: ; preds = %bb
- call void @llvm.amdgcn.raw.ptr.buffer.store.f32(float undef, ptr
addrspace(8) undef, i32 0, i32 undef, i32 0)
+ call void @llvm.amdgcn.raw.ptr.buffer.store.f32(float undef, ptr
addrspace(8) poison, i32 0, i32 undef, i32 0)
br label %endif
endif: ; preds = %bb28, %bb
@@ -85,7 +85,7 @@ loop:
%tmp23phi = phi i32 [ %tmp23, %loop ], [ 0, %entry ]
%tmp23 = add nuw i32 %tmp23phi, 1
%tmp27 = icmp ult i32 %arg, %tmp23
- call void @llvm.amdgcn.raw.ptr.buffer.store.f32(float undef, ptr
addrspace(8) undef, i32 0, i32 undef, i32 0)
+ call void @llvm.amdgcn.raw.ptr.buffer.store.f32(float undef, ptr
addrspace(8) poison, i32 0, i32 undef, i32 0)
br i1 %tmp27, label %loop, label %loopexit
loopexit:
@@ -136,7 +136,7 @@ loop:
br i1 %tmp27, label %then, label %endif
then: ; preds = %bb
- call void @llvm.amdgcn.raw.ptr.buffer.store.f32(float undef, ptr
addrspace(8) undef, i32 0, i32 undef, i32 0)
+ call void @llvm.amdgcn.raw.ptr.buffer.store.f32(float undef, ptr
addrspace(8) poison, i32 0, i32 undef, i32 0)
br label %endif
endif: ; preds = %bb28, %bb
diff --git a/llvm/test/CodeGen/AMDGPU/lower-work-group-id-intrinsics-hsa.ll
b/llvm/test/CodeGen/AMDGPU/lower-work-group-id-intrinsics-hsa.ll
index 018cdee038b91..a62427b2ce922 100644
--- a/llvm/test/CodeGen/AMDGPU/lower-work-group-id-intrinsics-hsa.ll
+++ b/llvm/test/CodeGen/AMDGPU/lower-work-group-id-intrinsics-hsa.ll
@@ -61,7 +61,7 @@ define amdgpu_kernel void @workgroup_ids_kernel() {
%ielemx = insertelement <3 x i32> poison, i32 %idx, i64 0
%ielemy = insertelement <3 x i32> %ielemx, i32 %idy, i64 1
%ielemz = insertelement <3 x i32> %ielemy, i32 %idz, i64 2
- call void @llvm.amdgcn.raw.ptr.buffer.store.v3i32(<3 x i32> %ielemz, ptr
addrspace(8) undef, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.raw.ptr.buffer.store.v3i32(<3 x i32> %ielemz, ptr
addrspace(8) poison, i32 0, i32 0, i32 0)
ret void
}
diff --git a/llvm/test/CodeGen/AMDGPU/lower-work-group-id-intrinsics-pal.ll
b/llvm/test/CodeGen/AMDGPU/lower-work-group-id-intrinsics-pal.ll
index dcec7288686b1..52b1d5e6c5b17 100644
--- a/llvm/test/CodeGen/AMDGPU/lower-work-group-id-intrinsics-pal.ll
+++ b/llvm/test/CodeGen/AMDGPU/lower-work-group-id-intrinsics-pal.ll
@@ -68,7 +68,7 @@ define amdgpu_cs void @_amdgpu_cs_main() {
%ielemx = insertelement <3 x i32> poison, i32 %idx, i64 0
%ielemy = insertelement <3 x i32> %ielemx, i32 %idy, i64 1
%ielemz = insertelement <3 x i32> %ielemy, i32 %idz, i64 2
- call void @llvm.amdgcn.raw.ptr.buffer.store.v3i32(<3 x i32> %ielemz, ptr
addrspace(8) undef, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.raw.ptr.buffer.store.v3i32(<3 x i32> %ielemz, ptr
addrspace(8) poison, i32 0, i32 0, i32 0)
ret void
}
diff --git a/llvm/test/CodeGen/AMDGPU/merge-store-crash.ll
b/llvm/test/CodeGen/AMDGPU/merge-store-crash.ll
index 0809891aec72b..66a217cb7be45 100644
--- a/llvm/test/CodeGen/AMDGPU/merge-store-crash.ll
+++ b/llvm/test/CodeGen/AMDGPU/merge-store-crash.ll
@@ -26,7 +26,7 @@ main_body:
%tmp9 = insertelement <4 x i32> %tmp8, i32 %tmp7, i32 1
%tmp10 = insertelement <4 x i32> %tmp9, i32 poison, i32 2
%tmp11 = insertelement <4 x i32> %tmp10, i32 poison, i32 3
- call void @llvm.amdgcn.struct.ptr.tbuffer.store.v4i32(<4 x i32> %tmp11, ptr
addrspace(8) undef, i32 0, i32 0, i32 %arg, i32 78, i32 3) #2
+ call void @llvm.amdgcn.struct.ptr.tbuffer.store.v4i32(<4 x i32> %tmp11, ptr
addrspace(8) poison, i32 0, i32 0, i32 %arg, i32 78, i32 3) #2
ret void
}
diff --git a/llvm/test/CodeGen/AMDGPU/merge-store-usedef.ll
b/llvm/test/CodeGen/AMDGPU/merge-store-usedef.ll
index 88cc5334d5c73..ae4fd6674f466 100644
--- a/llvm/test/CodeGen/AMDGPU/merge-store-usedef.ll
+++ b/llvm/test/CodeGen/AMDGPU/merge-store-usedef.ll
@@ -10,7 +10,7 @@ define amdgpu_vs void @test1(i32 %v) #0 {
store i32 %v, ptr addrspace(3) null
- call void @llvm.amdgcn.raw.ptr.tbuffer.store.i32(i32 %v, ptr addrspace(8)
undef, i32 0, i32 0, i32 68, i32 1)
+ call void @llvm.amdgcn.raw.ptr.tbuffer.store.i32(i32 %v, ptr addrspace(8)
poison, i32 0, i32 0, i32 68, i32 1)
%w = load i32, ptr addrspace(3) null
store i32 %w, ptr addrspace(3) %p1
diff --git a/llvm/test/CodeGen/AMDGPU/required-export-priority.ll
b/llvm/test/CodeGen/AMDGPU/required-export-priority.ll
index a37e2bf4eb294..f14cd4488ef1e 100644
--- a/llvm/test/CodeGen/AMDGPU/required-export-priority.ll
+++ b/llvm/test/CodeGen/AMDGPU/required-export-priority.ll
@@ -254,7 +254,7 @@ define amdgpu_ps void
@test_export_pos_before_param_across_load(i32 %idx) #0 {
; GCN-NEXT: s_endpgm
call void @llvm.amdgcn.exp.f32(i32 32, i32 15, float 1.0, float 1.0, float
1.0, float 1.0, i1 false, i1 false)
call void @llvm.amdgcn.exp.f32(i32 33, i32 15, float 1.0, float 1.0, float
1.0, float 0.5, i1 false, i1 false)
- %load = call float @llvm.amdgcn.raw.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx, i32 0, i32 0)
+ %load = call float @llvm.amdgcn.raw.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx, i32 0, i32 0)
call void @llvm.amdgcn.exp.f32(i32 12, i32 15, float 0.0, float 0.0, float
0.0, float %load, i1 true, i1 false)
ret void
}
diff --git a/llvm/test/CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll
b/llvm/test/CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll
index 096dfcc2c7667..4e6aa03d91876 100644
--- a/llvm/test/CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll
+++ b/llvm/test/CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll
@@ -318,7 +318,7 @@ define amdgpu_vs void
@reorder_local_load_tbuffer_store_local_load(ptr addrspace
%vdata = insertelement <4 x i32> poison, i32 %a1, i32 0
%vaddr.add = add i32 %vaddr, 32
- call void @llvm.amdgcn.struct.ptr.tbuffer.store.v4i32(<4 x i32> %vdata, ptr
addrspace(8) undef, i32 %vaddr.add, i32 0, i32 0, i32 228, i32 3)
+ call void @llvm.amdgcn.struct.ptr.tbuffer.store.v4i32(<4 x i32> %vdata, ptr
addrspace(8) poison, i32 %vaddr.add, i32 0, i32 0, i32 228, i32 3)
%tmp2 = load i32, ptr addrspace(3) %ptr2, align 4
diff --git a/llvm/test/CodeGen/AMDGPU/wave32.ll
b/llvm/test/CodeGen/AMDGPU/wave32.ll
index 94c4eebe32281..846e347a4947e 100644
--- a/llvm/test/CodeGen/AMDGPU/wave32.ll
+++ b/llvm/test/CodeGen/AMDGPU/wave32.ll
@@ -1959,7 +1959,7 @@ main_body:
br i1 %cc, label %endif, label %if
if:
- %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx, i32 0, i32 0, i32 0)
+ %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx, i32 0, i32 0, i32 0)
%out = fadd float %src, %src
%out.0 = call float @llvm.amdgcn.wwm.f32(float %out)
%out.1 = fadd float %src, %out.0
@@ -2046,7 +2046,7 @@ main_body:
br i1 %cc, label %endif, label %if
if:
- %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx, i32 0, i32 0, i32 0)
+ %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx, i32 0, i32 0, i32 0)
%out = fadd float %src, %src
%out.0 = call float @llvm.amdgcn.strict.wwm.f32(float %out)
%out.1 = fadd float %src, %out.0
@@ -2128,8 +2128,8 @@ define amdgpu_ps float @test_wqm2(i32 inreg %idx0, i32
inreg %idx1) #0 {
; GFX1064-NEXT: s_and_b64 exec, exec, s[2:3]
; GFX1064-NEXT: ; return to shader part epilog
main_body:
- %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx0, i32 0, i32 0, i32 0)
- %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx1, i32 0, i32 0, i32 0)
+ %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx0, i32 0, i32 0, i32 0)
+ %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx1, i32 0, i32 0, i32 0)
%out = fadd float %src0, %src1
%out.0 = bitcast float %out to i32
%out.1 = call i32 @llvm.amdgcn.wqm.i32(i32 %out.0)
diff --git a/llvm/test/CodeGen/AMDGPU/wqm.ll b/llvm/test/CodeGen/AMDGPU/wqm.ll
index 3845db2edf73f..abf690048bae3 100644
--- a/llvm/test/CodeGen/AMDGPU/wqm.ll
+++ b/llvm/test/CodeGen/AMDGPU/wqm.ll
@@ -125,7 +125,7 @@ main_body:
%tex.1 = bitcast <4 x float> %tex to <4 x i32>
%tex.2 = extractelement <4 x i32> %tex.1, i32 0
- call void @llvm.amdgcn.struct.ptr.buffer.store.v4f32(<4 x float> %tex, ptr
addrspace(8) undef, i32 %tex.2, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.v4f32(<4 x float> %tex, ptr
addrspace(8) poison, i32 %tex.2, i32 0, i32 0, i32 0)
ret <4 x float> %tex
}
@@ -247,7 +247,7 @@ define amdgpu_ps <4 x float> @test4_ptr_buf(<8 x i32> inreg
%rsrc, <4 x i32> inr
main_body:
%c.1 = mul i32 %c, %d
- call void @llvm.amdgcn.struct.ptr.buffer.store.v4f32(<4 x float> undef, ptr
addrspace(8) undef, i32 %c.1, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.v4f32(<4 x float> undef, ptr
addrspace(8) poison, i32 %c.1, i32 0, i32 0, i32 0)
%c.1.bc = bitcast i32 %c.1 to float
%tex = call <4 x float> @llvm.amdgcn.image.sample.1d.v4f32.f32(i32 15, float
%c.1.bc, <8 x i32> %rsrc, <4 x i32> %sampler, i1 false, i32 0, i32 0) #0
%tex0 = extractelement <4 x float> %tex, i32 0
@@ -327,8 +327,8 @@ define amdgpu_ps float @test5_ptr_buf(i32 inreg %idx0, i32
inreg %idx1) {
; GFX10-W32-NEXT: s_and_b32 exec_lo, exec_lo, s2
; GFX10-W32-NEXT: ; return to shader part epilog
main_body:
- %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx0, i32 0, i32 0, i32 0)
- %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx1, i32 0, i32 0, i32 0)
+ %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx0, i32 0, i32 0, i32 0)
+ %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx1, i32 0, i32 0, i32 0)
%out = fadd float %src0, %src1
%out.0 = call float @llvm.amdgcn.wqm.f32(float %out)
ret float %out.0
@@ -406,8 +406,8 @@ define amdgpu_ps float @test6_ptr_buf(i32 inreg %idx0, i32
inreg %idx1) {
; GFX10-W32-NEXT: s_and_b32 exec_lo, exec_lo, s2
; GFX10-W32-NEXT: ; return to shader part epilog
main_body:
- %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx0, i32 0, i32 0, i32 0)
- %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx1, i32 0, i32 0, i32 0)
+ %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx0, i32 0, i32 0, i32 0)
+ %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx1, i32 0, i32 0, i32 0)
%out = fadd float %src0, %src1
%out.0 = bitcast float %out to i32
%out.1 = call i32 @llvm.amdgcn.wqm.i32(i32 %out.0)
@@ -447,8 +447,8 @@ define amdgpu_ps float @test_wwm1(i32 inreg %idx0, i32
inreg %idx1) {
; GFX10-W32-NEXT: v_mov_b32_e32 v0, v1
; GFX10-W32-NEXT: ; return to shader part epilog
main_body:
- %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx0, i32 0, i32 0, i32 0)
- %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx1, i32 0, i32 0, i32 0)
+ %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx0, i32 0, i32 0, i32 0)
+ %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx1, i32 0, i32 0, i32 0)
%out = fadd float %src0, %src1
%out.0 = call float @llvm.amdgcn.wwm.f32(float %out)
ret float %out.0
@@ -484,8 +484,8 @@ define amdgpu_ps float @test_wwm2(i32 inreg %idx0, i32
inreg %idx1) {
; GFX10-W32-NEXT: v_mov_b32_e32 v0, v1
; GFX10-W32-NEXT: ; return to shader part epilog
main_body:
- %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx0, i32 0, i32 0, i32 0)
- %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx1, i32 0, i32 0, i32 0)
+ %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx0, i32 0, i32 0, i32 0)
+ %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx1, i32 0, i32 0, i32 0)
%src0.0 = bitcast float %src0 to i32
%src1.0 = bitcast float %src1 to i32
%out = add i32 %src0.0, %src1.0
@@ -549,7 +549,7 @@ main_body:
br i1 %cc, label %endif, label %if
if:
- %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx, i32 0, i32 0, i32 0)
+ %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx, i32 0, i32 0, i32 0)
%out = fadd float %src, %src
%out.0 = call float @llvm.amdgcn.wwm.f32(float %out)
%out.1 = fadd float %src, %out.0
@@ -612,7 +612,7 @@ main_body:
br i1 %cc, label %endif, label %if
if:
- %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx, i32 0, i32 0, i32 0)
+ %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx, i32 0, i32 0, i32 0)
%out = fadd float %src, %src
%out.0 = call float @llvm.amdgcn.wwm.f32(float %out)
br label %endif
@@ -666,9 +666,9 @@ define amdgpu_ps float @test_wwm5(i32 inreg %idx0, i32
inreg %idx1) {
; GFX10-W32-NEXT: s_and_b32 exec_lo, exec_lo, s2
; GFX10-W32-NEXT: ; return to shader part epilog
main_body:
- %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx0, i32 0, i32 0, i32 0)
- call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %src0, ptr
addrspace(8) undef, i32 %idx0, i32 0, i32 0, i32 0)
- %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx1, i32 0, i32 0, i32 0)
+ %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx0, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %src0, ptr
addrspace(8) poison, i32 %idx0, i32 0, i32 0, i32 0)
+ %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx1, i32 0, i32 0, i32 0)
%temp = fadd float %src1, %src1
%temp.0 = call float @llvm.amdgcn.wwm.f32(float %temp)
%out = fadd float %temp.0, %temp.0
@@ -857,13 +857,13 @@ define amdgpu_ps void @test_wwm_set_inactive1(i32 inreg
%idx) {
; GFX10-W32-NEXT: buffer_store_dword v2, v1, s[0:3], 0 idxen
; GFX10-W32-NEXT: s_endpgm
main_body:
- %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx, i32 0, i32 0, i32 0)
+ %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx, i32 0, i32 0, i32 0)
%src.0 = bitcast float %src to i32
%src.1 = call i32 @llvm.amdgcn.set.inactive.i32(i32 %src.0, i32 0)
%out = add i32 %src.1, %src.1
%out.0 = call i32 @llvm.amdgcn.wwm.i32(i32 %out)
%out.1 = bitcast i32 %out.0 to float
- call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %out.1, ptr
addrspace(8) undef, i32 %idx, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %out.1, ptr
addrspace(8) poison, i32 %idx, i32 0, i32 0, i32 0)
ret void
}
@@ -899,8 +899,8 @@ define amdgpu_ps float @test_strict_wqm1(i32 inreg %idx0,
i32 inreg %idx1) {
; GFX10-W32-NEXT: v_mov_b32_e32 v0, v1
; GFX10-W32-NEXT: ; return to shader part epilog
main_body:
- %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx0, i32 0, i32 0, i32 0)
- %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx1, i32 0, i32 0, i32 0)
+ %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx0, i32 0, i32 0, i32 0)
+ %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx1, i32 0, i32 0, i32 0)
%out = fadd float %src0, %src1
%out.0 = call float @llvm.amdgcn.strict.wqm.f32(float %out)
ret float %out.0
@@ -938,8 +938,8 @@ define amdgpu_ps float @test_strict_wqm2(i32 inreg %idx0,
i32 inreg %idx1) {
; GFX10-W32-NEXT: v_mov_b32_e32 v0, v1
; GFX10-W32-NEXT: ; return to shader part epilog
main_body:
- %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx0, i32 0, i32 0, i32 0)
- %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx1, i32 0, i32 0, i32 0)
+ %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx0, i32 0, i32 0, i32 0)
+ %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx1, i32 0, i32 0, i32 0)
%src0.0 = bitcast float %src0 to i32
%src1.0 = bitcast float %src1 to i32
%out = add i32 %src0.0, %src1.0
@@ -1005,7 +1005,7 @@ main_body:
br i1 %cc, label %endif, label %if
if:
- %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx, i32 0, i32 0, i32 0)
+ %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx, i32 0, i32 0, i32 0)
%out = fadd float %src, %src
%out.0 = call float @llvm.amdgcn.strict.wqm.f32(float %out)
%out.1 = fadd float %src, %out.0
@@ -1070,7 +1070,7 @@ main_body:
br i1 %cc, label %endif, label %if
if:
- %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx, i32 0, i32 0, i32 0)
+ %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx, i32 0, i32 0, i32 0)
%out = fadd float %src, %src
%out.0 = call float @llvm.amdgcn.strict.wqm.f32(float %out)
br label %endif
@@ -1127,9 +1127,9 @@ define amdgpu_ps float @test_strict_wqm5(i32 inreg %idx0,
i32 inreg %idx1) {
; GFX10-W32-NEXT: s_and_b32 exec_lo, exec_lo, s2
; GFX10-W32-NEXT: ; return to shader part epilog
main_body:
- %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx0, i32 0, i32 0, i32 0)
- call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %src0, ptr
addrspace(8) undef, i32 %idx0, i32 0, i32 0, i32 0)
- %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx1, i32 0, i32 0, i32 0)
+ %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx0, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %src0, ptr
addrspace(8) poison, i32 %idx0, i32 0, i32 0, i32 0)
+ %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx1, i32 0, i32 0, i32 0)
%temp = fadd float %src1, %src1
%temp.0 = call float @llvm.amdgcn.strict.wqm.f32(float %temp)
%out = fadd float %temp.0, %temp.0
@@ -1336,15 +1336,15 @@ define amdgpu_ps void @test_set_inactive2(i32 inreg
%idx0, i32 inreg %idx1) {
; GFX10-W32-NEXT: buffer_store_dword v1, v0, s[0:3], 0 idxen
; GFX10-W32-NEXT: s_endpgm
main_body:
- %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx1, i32 0, i32 0, i32 0)
+ %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx1, i32 0, i32 0, i32 0)
%src1.0 = bitcast float %src1 to i32
%src1.1 = call i32 @llvm.amdgcn.set.inactive.i32(i32 %src1.0, i32 undef)
- %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx0, i32 0, i32 0, i32 0)
+ %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx0, i32 0, i32 0, i32 0)
%src0.0 = bitcast float %src0 to i32
%src0.1 = call i32 @llvm.amdgcn.wqm.i32(i32 %src0.0)
%out = add i32 %src0.1, %src1.1
%out.0 = bitcast i32 %out to float
- call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %out.0, ptr
addrspace(8) undef, i32 %idx1, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %out.0, ptr
addrspace(8) poison, i32 %idx1, i32 0, i32 0, i32 0)
ret void
}
@@ -1420,7 +1420,7 @@ IF:
br label %END
ELSE:
- call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %data, ptr
addrspace(8) undef, i32 %c, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %data, ptr
addrspace(8) poison, i32 %c, i32 0, i32 0, i32 0)
br label %END
END:
@@ -1498,7 +1498,7 @@ IF:
br label %END
ELSE:
- call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %data, ptr
addrspace(8) undef, i32 %c, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %data, ptr
addrspace(8) poison, i32 %c, i32 0, i32 0, i32 0)
br label %END
END:
@@ -1570,16 +1570,16 @@ define amdgpu_ps <4 x float> @test_control_flow_2(<8 x
i32> inreg %rsrc, <4 x i3
main_body:
%idx.1 = extractelement <3 x i32> %idx, i32 0
%data.1 = extractelement <2 x float> %data, i32 0
- call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %data.1, ptr
addrspace(8) undef, i32 %idx.1, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %data.1, ptr
addrspace(8) poison, i32 %idx.1, i32 0, i32 0, i32 0)
; The load that determines the branch (and should therefore be WQM) is
; surrounded by stores that require disabled WQM.
%idx.2 = extractelement <3 x i32> %idx, i32 1
- %z = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx.2, i32 0, i32 0, i32 0)
+ %z = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx.2, i32 0, i32 0, i32 0)
%idx.3 = extractelement <3 x i32> %idx, i32 2
%data.3 = extractelement <2 x float> %data, i32 1
- call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %data.3, ptr
addrspace(8) undef, i32 %idx.3, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %data.3, ptr
addrspace(8) poison, i32 %idx.3, i32 0, i32 0, i32 0)
%cc = fcmp ogt float %z, 0.0
br i1 %cc, label %IF, label %ELSE
@@ -1671,7 +1671,7 @@ main_body:
%tex0 = extractelement <4 x float> %tex, i32 0
%dtex = call <4 x float> @llvm.amdgcn.image.sample.1d.v4f32.f32(i32 15,
float %tex0, <8 x i32> %rsrc, <4 x i32> %sampler, i1 false, i32 0, i32 0) #0
%dtex.1 = extractelement <4 x float> %dtex, i32 0
- call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %dtex.1, ptr
addrspace(8) undef, i32 %idx, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %dtex.1, ptr
addrspace(8) poison, i32 %idx, i32 0, i32 0, i32 0)
%cc = fcmp ogt float %dtex.1, 0.0
br i1 %cc, label %IF, label %ELSE
@@ -1741,8 +1741,8 @@ main_body:
br i1 %cond, label %IF, label %END
IF:
- %data = call float @llvm.amdgcn.raw.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 0, i32 0, i32 0)
- call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %data, ptr
addrspace(8) undef, i32 1, i32 0, i32 0, i32 0)
+ %data = call float @llvm.amdgcn.raw.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %data, ptr
addrspace(8) poison, i32 1, i32 0, i32 0, i32 0)
br label %END
END:
@@ -1821,14 +1821,14 @@ main_body:
%tex = call <4 x float> @llvm.amdgcn.image.sample.1d.v4f32.f32(i32 15, float
%coord, <8 x i32> %rsrc, <4 x i32> %sampler, i1 false, i32 0, i32 0) #0
%idx.0 = extractelement <2 x i32> %idx, i32 0
%data.0 = extractelement <2 x float> %data, i32 0
- call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %data.0, ptr
addrspace(8) undef, i32 %idx.0, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %data.0, ptr
addrspace(8) poison, i32 %idx.0, i32 0, i32 0, i32 0)
%z.cmp = fcmp olt float %z, 0.0
call void @llvm.amdgcn.kill(i1 %z.cmp)
%idx.1 = extractelement <2 x i32> %idx, i32 1
%data.1 = extractelement <2 x float> %data, i32 1
- call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %data.1, ptr
addrspace(8) undef, i32 %idx.1, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %data.1, ptr
addrspace(8) poison, i32 %idx.1, i32 0, i32 0, i32 0)
%tex2 = call <4 x float> @llvm.amdgcn.image.sample.1d.v4f32.f32(i32 15,
float %coord2, <8 x i32> %rsrc, <4 x i32> %sampler, i1 false, i32 0, i32 0) #0
%tex2.0 = extractelement <4 x float> %tex2, i32 0
%dtex = call <4 x float> @llvm.amdgcn.image.sample.1d.v4f32.f32(i32 15,
float %tex2.0, <8 x i32> %rsrc, <4 x i32> %sampler, i1 false, i32 0, i32 0) #0
@@ -1891,7 +1891,7 @@ main_body:
%tex0 = extractelement <4 x float> %tex, i32 0
%dtex = call <4 x float> @llvm.amdgcn.image.sample.1d.v4f32.f32(i32 15,
float %tex0, <8 x i32> %rsrc, <4 x i32> %sampler, i1 false, i32 0, i32 0) #0
- call void @llvm.amdgcn.raw.ptr.buffer.store.f32(float %data, ptr
addrspace(8) undef, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.raw.ptr.buffer.store.f32(float %data, ptr
addrspace(8) poison, i32 0, i32 0, i32 0)
%z.cmp = fcmp olt float %z, 0.0
call void @llvm.amdgcn.kill(i1 %z.cmp)
@@ -2071,17 +2071,17 @@ define amdgpu_ps void @test_alloca(float %data, i32 %a,
i32 %idx) nounwind {
entry:
%array = alloca [32 x i32], align 4, addrspace(5)
- call void @llvm.amdgcn.raw.ptr.buffer.store.f32(float %data, ptr
addrspace(8) undef, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.raw.ptr.buffer.store.f32(float %data, ptr
addrspace(8) poison, i32 0, i32 0, i32 0)
store volatile i32 %a, ptr addrspace(5) %array, align 4
- call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %data, ptr
addrspace(8) undef, i32 1, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %data, ptr
addrspace(8) poison, i32 1, i32 0, i32 0, i32 0)
%c.gep = getelementptr [32 x i32], ptr addrspace(5) %array, i32 0, i32 %idx
%c = load i32, ptr addrspace(5) %c.gep, align 4
%c.bc = bitcast i32 %c to float
%t = call <4 x float> @llvm.amdgcn.image.sample.1d.v4f32.f32(i32 15, float
%c.bc, <8 x i32> poison, <4 x i32> poison, i1 false, i32 0, i32 0) #0
- call void @llvm.amdgcn.raw.ptr.buffer.store.v4f32(<4 x float> %t, ptr
addrspace(8) undef, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.raw.ptr.buffer.store.v4f32(<4 x float> %t, ptr
addrspace(8) poison, i32 0, i32 0, i32 0)
ret void
}
@@ -2236,7 +2236,7 @@ else:
end:
%r = phi <4 x float> [ %r.if, %if ], [ %r.else, %else ]
- call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float 1.0, ptr
addrspace(8) undef, i32 %idx, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float 1.0, ptr
addrspace(8) poison, i32 %idx, i32 0, i32 0, i32 0)
ret <4 x float> %r
}
@@ -2348,8 +2348,8 @@ define amdgpu_ps float @test_strict_wwm1(i32 inreg %idx0,
i32 inreg %idx1) {
; GFX10-W32-NEXT: v_mov_b32_e32 v0, v1
; GFX10-W32-NEXT: ; return to shader part epilog
main_body:
- %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx0, i32 0, i32 0, i32 0)
- %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx1, i32 0, i32 0, i32 0)
+ %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx0, i32 0, i32 0, i32 0)
+ %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx1, i32 0, i32 0, i32 0)
%out = fadd float %src0, %src1
%out.0 = call float @llvm.amdgcn.strict.wwm.f32(float %out)
ret float %out.0
@@ -2385,8 +2385,8 @@ define amdgpu_ps float @test_strict_wwm2(i32 inreg %idx0,
i32 inreg %idx1) {
; GFX10-W32-NEXT: v_mov_b32_e32 v0, v1
; GFX10-W32-NEXT: ; return to shader part epilog
main_body:
- %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx0, i32 0, i32 0, i32 0)
- %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx1, i32 0, i32 0, i32 0)
+ %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx0, i32 0, i32 0, i32 0)
+ %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx1, i32 0, i32 0, i32 0)
%src0.0 = bitcast float %src0 to i32
%src1.0 = bitcast float %src1 to i32
%out = add i32 %src0.0, %src1.0
@@ -2450,7 +2450,7 @@ main_body:
br i1 %cc, label %endif, label %if
if:
- %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx, i32 0, i32 0, i32 0)
+ %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx, i32 0, i32 0, i32 0)
%out = fadd float %src, %src
%out.0 = call float @llvm.amdgcn.strict.wwm.f32(float %out)
%out.1 = fadd float %src, %out.0
@@ -2513,7 +2513,7 @@ main_body:
br i1 %cc, label %endif, label %if
if:
- %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx, i32 0, i32 0, i32 0)
+ %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx, i32 0, i32 0, i32 0)
%out = fadd float %src, %src
%out.0 = call float @llvm.amdgcn.strict.wwm.f32(float %out)
br label %endif
@@ -2567,9 +2567,9 @@ define amdgpu_ps float @test_strict_wwm5(i32 inreg %idx0,
i32 inreg %idx1) {
; GFX10-W32-NEXT: s_and_b32 exec_lo, exec_lo, s2
; GFX10-W32-NEXT: ; return to shader part epilog
main_body:
- %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx0, i32 0, i32 0, i32 0)
- call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %src0, ptr
addrspace(8) undef, i32 %idx0, i32 0, i32 0, i32 0)
- %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx1, i32 0, i32 0, i32 0)
+ %src0 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx0, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %src0, ptr
addrspace(8) poison, i32 %idx0, i32 0, i32 0, i32 0)
+ %src1 = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx1, i32 0, i32 0, i32 0)
%temp = fadd float %src1, %src1
%temp.0 = call float @llvm.amdgcn.strict.wwm.f32(float %temp)
%out = fadd float %temp.0, %temp.0
@@ -2754,13 +2754,13 @@ define amdgpu_ps void
@test_strict_wwm_set_inactive1(i32 inreg %idx) {
; GFX10-W32-NEXT: buffer_store_dword v2, v1, s[0:3], 0 idxen
; GFX10-W32-NEXT: s_endpgm
main_body:
- %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
undef, i32 %idx, i32 0, i32 0, i32 0)
+ %src = call float @llvm.amdgcn.struct.ptr.buffer.load.f32(ptr addrspace(8)
poison, i32 %idx, i32 0, i32 0, i32 0)
%src.0 = bitcast float %src to i32
%src.1 = call i32 @llvm.amdgcn.set.inactive.i32(i32 %src.0, i32 0)
%out = add i32 %src.1, %src.1
%out.0 = call i32 @llvm.amdgcn.strict.wwm.i32(i32 %out)
%out.1 = bitcast i32 %out.0 to float
- call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %out.1, ptr
addrspace(8) undef, i32 %idx, i32 0, i32 0, i32 0)
+ call void @llvm.amdgcn.struct.ptr.buffer.store.f32(float %out.1, ptr
addrspace(8) poison, i32 %idx, i32 0, i32 0, i32 0)
ret void
}
_______________________________________________
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
