RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-01-31 12:41:49 +01:00
Code Issues Projects Releases Wiki Activity

[SelectionDAG] Always preserve offset in MachinePointerInfo

Browse Source 
Previously, getWithOffset() would drop the offset if the base was null.
Because of this, MachineMemOperand would return the wrong result from
getAlign() in these cases.  MachineMemOperand stores the alignment of
the pointer without the offset.

A bunch of MIR tests changed because we print the offset now.

Split off from D77687.

Differential Revision: https://reviews.llvm.org/D78049
This commit is contained in:
Eli Friedman 2020-04-07 14:05:29 -07:00
parent 41f2e52f97
commit 50e1810fbf
18 changed files with 6637 additions and 6637 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
include/llvm/CodeGen/MachineMemOperand.h
View File

@ -58,8 +58,8 @@ struct MachinePointerInfo {
AddrSpace = v ? v->getAddressSpace() : 0;
}
explicit MachinePointerInfo(unsigned AddressSpace = 0)
: V((const Value *)nullptr), Offset(0), StackID(0),
explicit MachinePointerInfo(unsigned AddressSpace = 0, int64_t offset = 0)
: V((const Value *)nullptr), Offset(offset), StackID(0),
AddrSpace(AddressSpace) {}
explicit MachinePointerInfo(
@ -77,10 +77,10 @@ struct MachinePointerInfo {
MachinePointerInfo getWithOffset(int64_t O) const {
if (V.isNull())
return MachinePointerInfo(AddrSpace);
return MachinePointerInfo(AddrSpace, Offset + O);
if (V.is<const Value*>())
return MachinePointerInfo(V.get<const Value*>(), Offset+O, StackID);
return MachinePointerInfo(V.get<const PseudoSourceValue*>(), Offset+O,
return MachinePointerInfo(V.get<const Value*>(), Offset + O, StackID);
return MachinePointerInfo(V.get<const PseudoSourceValue*>(), Offset + O,
StackID);
}

6
test/CodeGen/AMDGPU/GlobalISel/legalize-load-constant-32bit.mir
View File

@ -14,13 +14,13 @@ body: |
; CI: [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[MV]](p4) :: (load 1, addrspace 6)
; CI: [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 1
; CI: [[PTR_ADD:%[0-9]+]]:_(p4) = G_PTR_ADD [[MV]], [[C1]](s64)
; CI: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p4) :: (load 1, addrspace 6)
; CI: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p4) :: (load 1 + 1, addrspace 6)
; CI: [[C2:%[0-9]+]]:_(s64) = G_CONSTANT i64 2
; CI: [[PTR_ADD1:%[0-9]+]]:_(p4) = G_PTR_ADD [[MV]], [[C2]](s64)
; CI: [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD1]](p4) :: (load 1, addrspace 6)
; CI: [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD1]](p4) :: (load 1 + 2, addrspace 6)
; CI: [[C3:%[0-9]+]]:_(s64) = G_CONSTANT i64 3
; CI: [[PTR_ADD2:%[0-9]+]]:_(p4) = G_PTR_ADD [[MV]], [[C3]](s64)
; CI: [[LOAD3:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD2]](p4) :: (load 1, addrspace 6)
; CI: [[LOAD3:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD2]](p4) :: (load 1 + 3, addrspace 6)
; CI: [[C4:%[0-9]+]]:_(s16) = G_CONSTANT i16 255
; CI: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[LOAD]](s32)
; CI: [[AND:%[0-9]+]]:_(s16) = G_AND [[TRUNC]], [[C4]]

2320
test/CodeGen/AMDGPU/GlobalISel/legalize-load-constant.mir
View File

File diff suppressed because it is too large Load Diff

2030
test/CodeGen/AMDGPU/GlobalISel/legalize-load-flat.mir
View File

File diff suppressed because it is too large Load Diff

1902
test/CodeGen/AMDGPU/GlobalISel/legalize-load-global.mir
View File

File diff suppressed because it is too large Load Diff

2356
test/CodeGen/AMDGPU/GlobalISel/legalize-load-local.mir
View File

File diff suppressed because it is too large Load Diff

2424
test/CodeGen/AMDGPU/GlobalISel/legalize-load-private.mir
View File

File diff suppressed because it is too large Load Diff

10
test/CodeGen/AMDGPU/GlobalISel/legalize-sextload-constant-32bit.mir
View File

@ -32,7 +32,7 @@ body: |
; CI: [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[MV]](p4) :: (load 2, addrspace 6)
; CI: [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 2
; CI: [[PTR_ADD:%[0-9]+]]:_(p4) = G_PTR_ADD [[MV]], [[C1]](s64)
; CI: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p4) :: (load 2, addrspace 6)
; CI: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p4) :: (load 2 + 2, addrspace 6)
; CI: [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535
; CI: [[COPY1:%[0-9]+]]:_(s32) = COPY [[LOAD]](s32)
; CI: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY1]], [[C2]]
@ -61,13 +61,13 @@ body: |
; CI: [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[MV]](p4) :: (load 1, addrspace 6)
; CI: [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 1
; CI: [[PTR_ADD:%[0-9]+]]:_(p4) = G_PTR_ADD [[MV]], [[C1]](s64)
; CI: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p4) :: (load 1, addrspace 6)
; CI: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p4) :: (load 1 + 1, addrspace 6)
; CI: [[C2:%[0-9]+]]:_(s64) = G_CONSTANT i64 2
; CI: [[PTR_ADD1:%[0-9]+]]:_(p4) = G_PTR_ADD [[MV]], [[C2]](s64)
; CI: [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD1]](p4) :: (load 1, addrspace 6)
; CI: [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD1]](p4) :: (load 1 + 2, addrspace 6)
; CI: [[C3:%[0-9]+]]:_(s64) = G_CONSTANT i64 3
; CI: [[PTR_ADD2:%[0-9]+]]:_(p4) = G_PTR_ADD [[MV]], [[C3]](s64)
; CI: [[LOAD3:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD2]](p4) :: (load 1, addrspace 6)
; CI: [[LOAD3:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD2]](p4) :: (load 1 + 3, addrspace 6)
; CI: [[C4:%[0-9]+]]:_(s16) = G_CONSTANT i16 255
; CI: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[LOAD]](s32)
; CI: [[AND:%[0-9]+]]:_(s16) = G_AND [[TRUNC]], [[C4]]
@ -149,7 +149,7 @@ body: |
; CI: [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[MV]](p4) :: (load 1, addrspace 6)
; CI: [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 1
; CI: [[PTR_ADD:%[0-9]+]]:_(p4) = G_PTR_ADD [[MV]], [[C1]](s64)
; CI: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p4) :: (load 1, addrspace 6)
; CI: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p4) :: (load 1 + 1, addrspace 6)
; CI: [[C2:%[0-9]+]]:_(s16) = G_CONSTANT i16 255
; CI: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[LOAD]](s32)
; CI: [[AND:%[0-9]+]]:_(s16) = G_AND [[TRUNC]], [[C2]]

2044
test/CodeGen/AMDGPU/GlobalISel/legalize-store-global.mir
View File

File diff suppressed because it is too large Load Diff

4
test/CodeGen/AMDGPU/GlobalISel/legalize-store.mir
View File

@ -148,7 +148,7 @@ body: |
; SI: G_STORE [[EXTRACT]](<2 x s32>), [[COPY]](p1) :: (store 8, align 4, addrspace 1)
; SI: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 8
; SI: [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[COPY]], [[C]](s64)
; SI: G_STORE [[EXTRACT1]](s32), [[PTR_ADD]](p1) :: (store 4, addrspace 1)
; SI: G_STORE [[EXTRACT1]](s32), [[PTR_ADD]](p1) :: (store 4 + 8, addrspace 1)
; VI-LABEL: name: test_store_global_v3s32
; VI: [[COPY:%[0-9]+]]:_(p1) = COPY $vgpr0_vgpr1
; VI: [[COPY1:%[0-9]+]]:_(<3 x s32>) = COPY $vgpr2_vgpr3_vgpr4
@ -332,7 +332,7 @@ body: |
; SI: G_STORE [[EXTRACT]](s64), [[COPY1]](p1) :: (store 8, align 16, addrspace 1)
; SI: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 8
; SI: [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[COPY1]], [[C]](s64)
; SI: G_STORE [[EXTRACT1]](s32), [[PTR_ADD]](p1) :: (store 4, align 8, addrspace 1)
; SI: G_STORE [[EXTRACT1]](s32), [[PTR_ADD]](p1) :: (store 4 + 8, align 8, addrspace 1)
; VI-LABEL: name: test_store_global_96
; VI: [[COPY:%[0-9]+]]:_(s96) = COPY $vgpr0_vgpr1_vgpr2
; VI: [[COPY1:%[0-9]+]]:_(p1) = COPY $vgpr3_vgpr4

10
test/CodeGen/AMDGPU/GlobalISel/legalize-zextload-constant-32bit.mir
View File

@ -32,7 +32,7 @@ body: |
; CI: [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[MV]](p4) :: (load 2, addrspace 6)
; CI: [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 2
; CI: [[PTR_ADD:%[0-9]+]]:_(p4) = G_PTR_ADD [[MV]], [[C1]](s64)
; CI: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p4) :: (load 2, addrspace 6)
; CI: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p4) :: (load 2 + 2, addrspace 6)
; CI: [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535
; CI: [[COPY1:%[0-9]+]]:_(s32) = COPY [[LOAD]](s32)
; CI: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY1]], [[C2]]
@ -61,13 +61,13 @@ body: |
; CI: [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[MV]](p4) :: (load 1, addrspace 6)
; CI: [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 1
; CI: [[PTR_ADD:%[0-9]+]]:_(p4) = G_PTR_ADD [[MV]], [[C1]](s64)
; CI: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p4) :: (load 1, addrspace 6)
; CI: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p4) :: (load 1 + 1, addrspace 6)
; CI: [[C2:%[0-9]+]]:_(s64) = G_CONSTANT i64 2
; CI: [[PTR_ADD1:%[0-9]+]]:_(p4) = G_PTR_ADD [[MV]], [[C2]](s64)
; CI: [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD1]](p4) :: (load 1, addrspace 6)
; CI: [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD1]](p4) :: (load 1 + 2, addrspace 6)
; CI: [[C3:%[0-9]+]]:_(s64) = G_CONSTANT i64 3
; CI: [[PTR_ADD2:%[0-9]+]]:_(p4) = G_PTR_ADD [[MV]], [[C3]](s64)
; CI: [[LOAD3:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD2]](p4) :: (load 1, addrspace 6)
; CI: [[LOAD3:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD2]](p4) :: (load 1 + 3, addrspace 6)
; CI: [[C4:%[0-9]+]]:_(s16) = G_CONSTANT i16 255
; CI: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[LOAD]](s32)
; CI: [[AND:%[0-9]+]]:_(s16) = G_AND [[TRUNC]], [[C4]]
@ -151,7 +151,7 @@ body: |
; CI: [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[MV]](p4) :: (load 1, addrspace 6)
; CI: [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 1
; CI: [[PTR_ADD:%[0-9]+]]:_(p4) = G_PTR_ADD [[MV]], [[C1]](s64)
; CI: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p4) :: (load 1, addrspace 6)
; CI: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p4) :: (load 1 + 1, addrspace 6)
; CI: [[C2:%[0-9]+]]:_(s16) = G_CONSTANT i16 255
; CI: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[LOAD]](s32)
; CI: [[AND:%[0-9]+]]:_(s16) = G_AND [[TRUNC]], [[C2]]

56
test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.s.buffer.load.ll
View File

@ -1803,8 +1803,8 @@ define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset(<4 x i32> inreg
; GFX6: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 16, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 48, align 4)
; GFX6: [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
; GFX6: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
; GFX6: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
@ -1851,8 +1851,8 @@ define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset(<4 x i32> inreg
; GFX7: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 16, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 48, align 4)
; GFX7: [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
; GFX7: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
; GFX7: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
@ -1899,8 +1899,8 @@ define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset(<4 x i32> inreg
; GFX8: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 16, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 48, align 4)
; GFX8: [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
; GFX8: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
; GFX8: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
@ -2281,8 +2281,8 @@ define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4032(<4 x i3
; GFX6: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4032, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4048, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 16, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 48, align 4)
; GFX6: [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
; GFX6: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
; GFX6: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
@ -2329,8 +2329,8 @@ define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4032(<4 x i3
; GFX7: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4032, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4048, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 16, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 48, align 4)
; GFX7: [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
; GFX7: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
; GFX7: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
@ -2377,8 +2377,8 @@ define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4032(<4 x i3
; GFX8: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4032, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4048, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 16, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 48, align 4)
; GFX8: [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
; GFX8: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
; GFX8: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
@ -2431,8 +2431,8 @@ define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4036(<4 x i3
; GFX6: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4036
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 16, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 48, align 4)
; GFX6: [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
; GFX6: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
; GFX6: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
@ -2479,8 +2479,8 @@ define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4036(<4 x i3
; GFX7: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4036
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 16, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 32, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 16, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 48, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 48, align 4)
; GFX7: [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
; GFX7: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
; GFX7: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
@ -2527,8 +2527,8 @@ define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4036(<4 x i3
; GFX8: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4032, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4048, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4064, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 16, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4080, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 48, align 4)
; GFX8: [[REG_SEQUENCE1:%[0-9]+]]:vreg_512 = REG_SEQUENCE [[BUFFER_LOAD_DWORDX4_OFFEN]], %subreg.sub0_sub1_sub2_sub3, [[BUFFER_LOAD_DWORDX4_OFFEN1]], %subreg.sub4_sub5_sub6_sub7, [[BUFFER_LOAD_DWORDX4_OFFEN2]], %subreg.sub8_sub9_sub10_sub11, [[BUFFER_LOAD_DWORDX4_OFFEN3]], %subreg.sub12_sub13_sub14_sub15
; GFX8: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub0
; GFX8: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[REG_SEQUENCE1]].sub1
@ -2929,7 +2929,7 @@ define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4095(<4 x i32> %rsrc)
; GFX6: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[REG_SEQUENCE2]], [[COPY5]], implicit $exec
; GFX6: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_1]], [[V_CMP_EQ_U64_e64_]], implicit-def $scc
; GFX6: [[REG_SEQUENCE3:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX6: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4095, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 4, align 1)
; GFX6: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4095, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 4 + 4095, align 1)
; GFX6: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX6: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX6: S_CBRANCH_EXECNZ %bb.2, implicit $exec
@ -2962,7 +2962,7 @@ define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4095(<4 x i32> %rsrc)
; GFX7: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[REG_SEQUENCE2]], [[COPY5]], implicit $exec
; GFX7: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_1]], [[V_CMP_EQ_U64_e64_]], implicit-def $scc
; GFX7: [[REG_SEQUENCE3:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX7: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4095, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 4, align 1)
; GFX7: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4095, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 4 + 4095, align 1)
; GFX7: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX7: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX7: S_CBRANCH_EXECNZ %bb.2, implicit $exec
@ -2995,7 +2995,7 @@ define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4095(<4 x i32> %rsrc)
; GFX8: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[REG_SEQUENCE2]], [[COPY5]], implicit $exec
; GFX8: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_1]], [[V_CMP_EQ_U64_e64_]], implicit-def $scc
; GFX8: [[REG_SEQUENCE3:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX8: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4095, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 4, align 1)
; GFX8: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4095, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 4 + 4095, align 1)
; GFX8: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX8: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX8: S_CBRANCH_EXECNZ %bb.2, implicit $exec
@ -3104,7 +3104,7 @@ define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4096(<4 x i32> %rsrc)
; GFX8: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[REG_SEQUENCE2]], [[COPY5]], implicit $exec
; GFX8: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_1]], [[V_CMP_EQ_U64_e64_]], implicit-def $scc
; GFX8: [[REG_SEQUENCE3:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX8: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4095, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 4)
; GFX8: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4095, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 4 + 4096)
; GFX8: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX8: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX8: S_CBRANCH_EXECNZ %bb.2, implicit $exec
@ -4115,8 +4115,8 @@ define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4
; GFX6: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[REG_SEQUENCE2]], [[COPY5]], implicit $exec
; GFX6: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_1]], [[V_CMP_EQ_U64_e64_]], implicit-def $scc
; GFX6: [[REG_SEQUENCE3:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4064, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4080, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4064, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 4064, align 4)
; GFX6: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4080, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 4064, align 4)
; GFX6: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX6: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX6: S_CBRANCH_EXECNZ %bb.2, implicit $exec
@ -4165,8 +4165,8 @@ define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4
; GFX7: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[REG_SEQUENCE2]], [[COPY5]], implicit $exec
; GFX7: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_1]], [[V_CMP_EQ_U64_e64_]], implicit-def $scc
; GFX7: [[REG_SEQUENCE3:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4064, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4080, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4064, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 4064, align 4)
; GFX7: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4080, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 4064, align 4)
; GFX7: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX7: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX7: S_CBRANCH_EXECNZ %bb.2, implicit $exec
@ -4215,8 +4215,8 @@ define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4
; GFX8: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[REG_SEQUENCE2]], [[COPY5]], implicit $exec
; GFX8: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_1]], [[V_CMP_EQ_U64_e64_]], implicit-def $scc
; GFX8: [[REG_SEQUENCE3:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4064, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4080, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4064, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 4064, align 4)
; GFX8: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE3]], [[S_MOV_B32_]], 4080, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load 16 + 4064, align 4)
; GFX8: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX8: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX8: S_CBRANCH_EXECNZ %bb.2, implicit $exec

34
test/CodeGen/AMDGPU/GlobalISel/regbankselect-amdgcn.s.buffer.load.ll
View File

@ -308,8 +308,8 @@ define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset(<4 x i32> inreg
; CHECK: [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
; CHECK: [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load 16 + 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load 16 + 48, align 4)
; CHECK: [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<16 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>), [[AMDGPU_BUFFER_LOAD2]](<4 x s32>), [[AMDGPU_BUFFER_LOAD3]](<4 x s32>)
; CHECK: [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32), [[UV4:%[0-9]+]]:vgpr(s32), [[UV5:%[0-9]+]]:vgpr(s32), [[UV6:%[0-9]+]]:vgpr(s32), [[UV7:%[0-9]+]]:vgpr(s32), [[UV8:%[0-9]+]]:vgpr(s32), [[UV9:%[0-9]+]]:vgpr(s32), [[UV10:%[0-9]+]]:vgpr(s32), [[UV11:%[0-9]+]]:vgpr(s32), [[UV12:%[0-9]+]]:vgpr(s32), [[UV13:%[0-9]+]]:vgpr(s32), [[UV14:%[0-9]+]]:vgpr(s32), [[UV15:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<16 x s32>)
; CHECK: $vgpr0 = COPY [[UV]](s32)
@ -399,8 +399,8 @@ define amdgpu_ps void @s_buffer_load_i512_vgpr_offset(<4 x i32> inreg %rsrc, i32
; CHECK: [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
; CHECK: [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load 16 + 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(s128) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load 16 + 48, align 4)
; CHECK: [[MV:%[0-9]+]]:vgpr(s512) = G_MERGE_VALUES [[AMDGPU_BUFFER_LOAD]](s128), [[AMDGPU_BUFFER_LOAD1]](s128), [[AMDGPU_BUFFER_LOAD2]](s128), [[AMDGPU_BUFFER_LOAD3]](s128)
; CHECK: [[UV:%[0-9]+]]:vgpr(s128), [[UV1:%[0-9]+]]:vgpr(s128), [[UV2:%[0-9]+]]:vgpr(s128), [[UV3:%[0-9]+]]:vgpr(s128) = G_UNMERGE_VALUES [[MV]](s512)
; CHECK: G_STORE [[UV]](s128), [[DEF]](p1) :: (store 16 into `i512 addrspace(1)* undef`, align 8, addrspace 1)
@ -463,8 +463,8 @@ define amdgpu_ps void @s_buffer_load_v32i16_vgpr_offset(<4 x i32> inreg %rsrc, i
; CHECK: [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
; CHECK: [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load 16 + 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<8 x s16>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load 16 + 48, align 4)
; CHECK: [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<32 x s16>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<8 x s16>), [[AMDGPU_BUFFER_LOAD1]](<8 x s16>), [[AMDGPU_BUFFER_LOAD2]](<8 x s16>), [[AMDGPU_BUFFER_LOAD3]](<8 x s16>)
; CHECK: [[UV:%[0-9]+]]:vgpr(<8 x s16>), [[UV1:%[0-9]+]]:vgpr(<8 x s16>), [[UV2:%[0-9]+]]:vgpr(<8 x s16>), [[UV3:%[0-9]+]]:vgpr(<8 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<32 x s16>)
; CHECK: G_STORE [[UV]](<8 x s16>), [[DEF]](p1) :: (store 16 into `<32 x i16> addrspace(1)* undef`, align 64, addrspace 1)
@ -527,8 +527,8 @@ define amdgpu_ps void @s_buffer_load_v8i64_vgpr_offset(<4 x i32> inreg %rsrc, i3
; CHECK: [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
; CHECK: [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load 16 + 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<2 x s64>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load 16 + 48, align 4)
; CHECK: [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x s64>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<2 x s64>), [[AMDGPU_BUFFER_LOAD1]](<2 x s64>), [[AMDGPU_BUFFER_LOAD2]](<2 x s64>), [[AMDGPU_BUFFER_LOAD3]](<2 x s64>)
; CHECK: [[UV:%[0-9]+]]:vgpr(<2 x s64>), [[UV1:%[0-9]+]]:vgpr(<2 x s64>), [[UV2:%[0-9]+]]:vgpr(<2 x s64>), [[UV3:%[0-9]+]]:vgpr(<2 x s64>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<8 x s64>)
; CHECK: G_STORE [[UV]](<2 x s64>), [[DEF]](p1) :: (store 16 into `<8 x i64> addrspace(1)* undef`, align 64, addrspace 1)
@ -591,8 +591,8 @@ define amdgpu_ps void @s_buffer_load_v8p1_vgpr_offset(<4 x i32> inreg %rsrc, i32
; CHECK: [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
; CHECK: [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load 16 + 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<2 x p1>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load 16 + 48, align 4)
; CHECK: [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<8 x p1>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<2 x p1>), [[AMDGPU_BUFFER_LOAD1]](<2 x p1>), [[AMDGPU_BUFFER_LOAD2]](<2 x p1>), [[AMDGPU_BUFFER_LOAD3]](<2 x p1>)
; CHECK: [[UV:%[0-9]+]]:vgpr(<2 x p1>), [[UV1:%[0-9]+]]:vgpr(<2 x p1>), [[UV2:%[0-9]+]]:vgpr(<2 x p1>), [[UV3:%[0-9]+]]:vgpr(<2 x p1>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<8 x p1>)
; CHECK: G_STORE [[UV]](<2 x p1>), [[DEF]](p1) :: (store 16 into `<8 x i8 addrspace(1)*> addrspace(1)* undef`, align 64, addrspace 1)
@ -763,8 +763,8 @@ define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4032(<4 x i3
; CHECK: [[C2:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
; CHECK: [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4032, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4048, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4064, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4080, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4064, 0, 0 :: (dereferenceable invariant load 16 + 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C2]](s32), [[COPY4]], [[C1]], 4080, 0, 0 :: (dereferenceable invariant load 16 + 48, align 4)
; CHECK: [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<16 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>), [[AMDGPU_BUFFER_LOAD2]](<4 x s32>), [[AMDGPU_BUFFER_LOAD3]](<4 x s32>)
; CHECK: [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32), [[UV4:%[0-9]+]]:vgpr(s32), [[UV5:%[0-9]+]]:vgpr(s32), [[UV6:%[0-9]+]]:vgpr(s32), [[UV7:%[0-9]+]]:vgpr(s32), [[UV8:%[0-9]+]]:vgpr(s32), [[UV9:%[0-9]+]]:vgpr(s32), [[UV10:%[0-9]+]]:vgpr(s32), [[UV11:%[0-9]+]]:vgpr(s32), [[UV12:%[0-9]+]]:vgpr(s32), [[UV13:%[0-9]+]]:vgpr(s32), [[UV14:%[0-9]+]]:vgpr(s32), [[UV15:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<16 x s32>)
; CHECK: $vgpr0 = COPY [[UV]](s32)
@ -805,8 +805,8 @@ define amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4036(<4 x i3
; CHECK: [[C1:%[0-9]+]]:vgpr(s32) = G_CONSTANT i32 0
; CHECK: [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 0, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 16, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD2:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 32, 0, 0 :: (dereferenceable invariant load 16 + 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD3:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR]](<4 x s32>), [[C1]](s32), [[COPY4]], [[C]], 48, 0, 0 :: (dereferenceable invariant load 16 + 48, align 4)
; CHECK: [[CONCAT_VECTORS:%[0-9]+]]:vgpr(<16 x s32>) = G_CONCAT_VECTORS [[AMDGPU_BUFFER_LOAD]](<4 x s32>), [[AMDGPU_BUFFER_LOAD1]](<4 x s32>), [[AMDGPU_BUFFER_LOAD2]](<4 x s32>), [[AMDGPU_BUFFER_LOAD3]](<4 x s32>)
; CHECK: [[UV:%[0-9]+]]:vgpr(s32), [[UV1:%[0-9]+]]:vgpr(s32), [[UV2:%[0-9]+]]:vgpr(s32), [[UV3:%[0-9]+]]:vgpr(s32), [[UV4:%[0-9]+]]:vgpr(s32), [[UV5:%[0-9]+]]:vgpr(s32), [[UV6:%[0-9]+]]:vgpr(s32), [[UV7:%[0-9]+]]:vgpr(s32), [[UV8:%[0-9]+]]:vgpr(s32), [[UV9:%[0-9]+]]:vgpr(s32), [[UV10:%[0-9]+]]:vgpr(s32), [[UV11:%[0-9]+]]:vgpr(s32), [[UV12:%[0-9]+]]:vgpr(s32), [[UV13:%[0-9]+]]:vgpr(s32), [[UV14:%[0-9]+]]:vgpr(s32), [[UV15:%[0-9]+]]:vgpr(s32) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<16 x s32>)
; CHECK: $vgpr0 = COPY [[UV]](s32)
@ -1003,7 +1003,7 @@ define amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4095(<4 x i32> %rsrc)
; CHECK: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[MV1]](s64), [[UV1]](s64), implicit $exec
; CHECK: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_1]], [[V_CMP_EQ_U64_e64_]], implicit-def $scc
; CHECK: [[BUILD_VECTOR1:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[V_READFIRSTLANE_B32_]](s32), [[V_READFIRSTLANE_B32_1]](s32), [[V_READFIRSTLANE_B32_2]](s32), [[V_READFIRSTLANE_B32_3]](s32)
; CHECK: [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4095, 0, 0 :: (dereferenceable invariant load 4, align 1)
; CHECK: [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(s32) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4095, 0, 0 :: (dereferenceable invariant load 4 + 4095, align 1)
; CHECK: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK: S_CBRANCH_EXECNZ %bb.2, implicit $exec
@ -1448,8 +1448,8 @@ define amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4
; CHECK: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[MV1]](s64), [[UV1]](s64), implicit $exec
; CHECK: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_1]], [[V_CMP_EQ_U64_e64_]], implicit-def $scc
; CHECK: [[BUILD_VECTOR1:%[0-9]+]]:sgpr(<4 x s32>) = G_BUILD_VECTOR [[V_READFIRSTLANE_B32_]](s32), [[V_READFIRSTLANE_B32_1]](s32), [[V_READFIRSTLANE_B32_2]](s32), [[V_READFIRSTLANE_B32_3]](s32)
; CHECK: [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4064, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4080, 0, 0 :: (dereferenceable invariant load 16, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4064, 0, 0 :: (dereferenceable invariant load 16 + 4064, align 4)
; CHECK: [[AMDGPU_BUFFER_LOAD1:%[0-9]+]]:vgpr(<4 x s32>) = G_AMDGPU_BUFFER_LOAD [[BUILD_VECTOR1]](<4 x s32>), [[C3]](s32), [[C1]], [[C2]], 4080, 0, 0 :: (dereferenceable invariant load 16 + 4064, align 4)
; CHECK: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK: S_CBRANCH_EXECNZ %bb.2, implicit $exec

12
test/CodeGen/ARM/GlobalISel/arm-legalize-load-store.mir
View File

@ -110,9 +110,9 @@ body: |
; CHECK-NEXT: [[OFF:%[0-9]+]]:_(s32) = G_CONSTANT i32 4
; CHECK-NEXT: [[ADDR2:%[0-9]+]]:_(p0) = G_PTR_ADD [[ADDR1]], [[OFF]]
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(p0) = COPY [[ADDR2]]
; CHECK-NEXT: [[V2:%[0-9]+]]:_(s32) = G_LOAD [[COPY]](p0) :: (load 4, align 1)
; CHECK-NEXT: [[V2:%[0-9]+]]:_(s32) = G_LOAD [[COPY]](p0) :: (load 4 + 4, align 1)
; CHECK-NEXT: G_STORE [[V1]](s32), [[ADDR1]](p0) :: (store 4, align 1)
; CHECK-NEXT: G_STORE [[V2]](s32), [[ADDR2]](p0) :: (store 4, align 1)
; CHECK-NEXT: G_STORE [[V2]](s32), [[ADDR2]](p0) :: (store 4 + 4, align 1)
%0(p0) = COPY $r0
%1(s64) = G_LOAD %0(p0) :: (load 8, align 1)
G_STORE %1(s64), %0(p0) :: (store 8, align 1)
@ -147,19 +147,19 @@ body: |
; CHECK-NEXT: [[OFF:%[0-9]+]]:_(s32) = G_CONSTANT i32 4
; CHECK-NEXT: [[ADDR2:%[0-9]+]]:_(p0) = G_PTR_ADD [[ADDR1]], [[OFF]]
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(p0) = COPY [[ADDR2]]
; CHECK-NEXT: [[V2:%[0-9]+]]:_(s32) = G_LOAD [[COPY]](p0) :: (load 4, align 1)
; CHECK-NEXT: [[V2:%[0-9]+]]:_(s32) = G_LOAD [[COPY]](p0) :: (load 4 + 4, align 1)
; CHECK-NEXT: G_STORE [[V1]](s32), [[ADDR1]](p0) :: (store 4, align 1)
; CHECK-NEXT: [[COPY2:%[0-9]+]]:_(p0) = COPY [[ADDR2]]
; CHECK-NEXT: G_STORE [[V2]](s32), [[COPY2]](p0) :: (store 4, align 1)
; CHECK-NEXT: G_STORE [[V2]](s32), [[COPY2]](p0) :: (store 4 + 4, align 1)
%0(p0) = COPY $r0
%1(s64) = G_LOAD %0(p0) :: (load 8, align 1)
G_STORE %1(s64), %0(p0) :: (store 8, align 1)
; CHECK: [[V1:%[0-9]+]]:_(s32) = G_LOAD [[ADDR1]](p0) :: (load 4)
; CHECK-NEXT: [[COPY3:%[0-9]+]]:_(p0) = COPY [[ADDR2]]
; CHECK-NEXT: [[V2:%[0-9]+]]:_(s32) = G_LOAD [[COPY3]](p0) :: (load 4)
; CHECK-NEXT: [[V2:%[0-9]+]]:_(s32) = G_LOAD [[COPY3]](p0) :: (load 4 + 4)
; CHECK-NEXT: G_STORE [[V1]](s32), [[ADDR1]](p0) :: (store 4)
; CHECK-NEXT: G_STORE [[V2]](s32), [[ADDR2]](p0) :: (store 4)
; CHECK-NEXT: G_STORE [[V2]](s32), [[ADDR2]](p0) :: (store 4 + 4)
%2(s64) = G_LOAD %0(p0) :: (load 8, align 4)
G_STORE %2(s64), %0(p0) :: (store 8, align 4)

46
test/CodeGen/PowerPC/aix-cc-abi.ll
View File

@ -1842,23 +1842,23 @@ entry:
; 32BIT-DAG: renamable $r[[SCRATCHREG:[0-9]+]] = LWZtoc %const.10, $r2 :: (load 4 from got)
; 32BIT-DAG: renamable $r[[SCRATCHREG:[0-9]+]] = LWZtoc %const.11, $r2 :: (load 4 from got)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 56, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 60, $r1 :: (store 4)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 60, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 64, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 68, $r1 :: (store 4)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 68, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 72, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 76, $r1 :: (store 4)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 76, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 80, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW renamable $r[[SCRATCHREG:[0-9]+]], 84, $r1 :: (store 4)
; 32BIT-DAG: STW renamable $r[[SCRATCHREG:[0-9]+]], 84, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 88, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 92, $r1 :: (store 4)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 92, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 96, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 100, $r1 :: (store 4)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 100, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 104, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 108, $r1 :: (store 4)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 108, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 112, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 116, $r1 :: (store 4)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 116, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 120, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 124, $r1 :: (store 4)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 124, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[SCRATCHREG:[0-9]+]], 128, $r1 :: (store 4)
; 32BIT-DAG: renamable $r[[REGF1:[0-9]+]] = LWZtoc @f14, $r2 :: (load 4 from got)
; 32BIT-DAG: renamable $r3 = LWZ 0, killed renamable $r[[REGF1]] :: (load 4 from @f14)
@ -2243,33 +2243,33 @@ define void @caller_mix() {
; 32BIT-DAG: $r9 = LI 7
; 32BIT-DAG: $r10 = LI 8
; 32BIT-DAG: STW killed renamable $r[[REG1:[0-9]+]], 56, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW renamable $r[[REG2:[0-9]+]], 60, $r1 :: (store 4)
; 32BIT-DAG: STW renamable $r[[REG2:[0-9]+]], 60, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[REG3:[0-9]+]], 64, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW renamable $r[[REG4:[0-9]+]], 68, $r1 :: (store 4)
; 32BIT-DAG: STW renamable $r[[REG4:[0-9]+]], 68, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[REG5:[0-9]+]], 72, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW renamable $r[[REG6:[0-9]+]], 76, $r1 :: (store 4)
; 32BIT-DAG: STW renamable $r[[REG6:[0-9]+]], 76, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[REG7:[0-9]+]], 80, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW renamable $r[[REG8:[0-9]+]], 84, $r1 :: (store 4)
; 32BIT-DAG: STW renamable $r[[REG8:[0-9]+]], 84, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[REG9:[0-9]+]], 88, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW renamable $r[[REG10:[0-9]+]], 92, $r1 :: (store 4)
; 32BIT-DAG: STW renamable $r[[REG10:[0-9]+]], 92, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[REG11:[0-9]+]], 96, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW renamable $r[[REG12:[0-9]+]], 100, $r1 :: (store 4)
; 32BIT-DAG: STW renamable $r[[REG12:[0-9]+]], 100, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[REG13:[0-9]+]], 104, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW renamable $r[[REG14:[0-9]+]], 108, $r1 :: (store 4)
; 32BIT-DAG: STW renamable $r[[REG14:[0-9]+]], 108, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[REG15:[0-9]+]], 112, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW renamable $r[[REG16:[0-9]+]], 116, $r1 :: (store 4)
; 32BIT-DAG: STW renamable $r[[REG16:[0-9]+]], 116, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[REG17:[0-9]+]], 120, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW killed renamable $r[[REG18:[0-9]+]], 128, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW renamable $r[[REG19:[0-9]+]], 124, $r1 :: (store 4)
; 32BIT-DAG: STW killed renamable $r[[REG20:[0-9]+]], 132, $r1 :: (store 4)
; 32BIT-DAG: STW renamable $r[[REG19:[0-9]+]], 124, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[REG20:[0-9]+]], 132, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[REG21:[0-9]+]], 136, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW killed renamable $r[[REG22:[0-9]+]], 140, $r1 :: (store 4)
; 32BIT-DAG: STW killed renamable $r[[REG22:[0-9]+]], 140, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[REG23:[0-9]+]], 144, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW killed renamable $r[[REG24:[0-9]+]], 148, $r1 :: (store 4)
; 32BIT-DAG: STW killed renamable $r[[REG24:[0-9]+]], 148, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[REG25:[0-9]+]], 152, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW killed renamable $r[[REG26:[0-9]+]], 156, $r1 :: (store 4)
; 32BIT-DAG: STW killed renamable $r[[REG26:[0-9]+]], 156, $r1 :: (store 4 + 4)
; 32BIT-DAG: STW killed renamable $r[[REG27:[0-9]+]], 160, $r1 :: (store 4, align 8)
; 32BIT-DAG: STW killed renamable $r[[REG28:[0-9]+]], 164, $r1 :: (store 4)
; 32BIT-DAG: STW killed renamable $r[[REG28:[0-9]+]], 164, $r1 :: (store 4 + 4)
; 32BIT-NEXT: BL_NOP <mcsymbol .mix_floats>, csr_aix32, implicit-def dead $lr, implicit $rm, implicit $r3, implicit $r4, implicit $r5, implicit $r6, implicit $r7, implicit $r8, implicit $r9, implicit $r10, implicit $f1, implicit $f2, implicit $f3, implicit $f4, implicit $f5, implicit $f6, implicit $f7, implicit $f8, implicit $f9, implicit $f10, implicit $f11, implicit $f12, implicit $f13, implicit $r2, implicit-def $r1, implicit-def dead $r3
; 32BIT-NEXT: ADJCALLSTACKUP 168, 0, implicit-def dead $r1, implicit $r1

4
test/CodeGen/X86/GlobalISel/legalize-memop-scalar-32.mir
View File

@ -50,10 +50,10 @@ body: |
; X32: [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p0) :: (load 4, align 8)
; X32: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 4
; X32: [[GEP:%[0-9]+]]:_(p0) = G_PTR_ADD [[DEF]], [[C]](s32)
; X32: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[GEP]](p0) :: (load 4)
; X32: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[GEP]](p0) :: (load 4 + 4)
; X32: G_STORE [[LOAD]](s32), [[DEF]](p0) :: (store 4, align 8)
; X32: [[GEP1:%[0-9]+]]:_(p0) = G_PTR_ADD [[DEF]], [[C]](s32)
; X32: G_STORE [[LOAD1]](s32), [[GEP1]](p0) :: (store 4)
; X32: G_STORE [[LOAD1]](s32), [[GEP1]](p0) :: (store 4 + 4)
%0:_(p0) = IMPLICIT_DEF
%1:_(s64) = G_LOAD %0 :: (load 8)

2
test/CodeGen/X86/GlobalISel/legalize-undef.mir
View File

@ -34,7 +34,7 @@ body: |
; X32: G_STORE [[DEF3]](s32), [[DEF]](p0) :: (store 4, align 8)
; X32: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 4
; X32: [[PTR_ADD:%[0-9]+]]:_(p0) = G_PTR_ADD [[DEF]], [[C1]](s32)
; X32: G_STORE [[DEF3]](s32), [[PTR_ADD]](p0) :: (store 4)
; X32: G_STORE [[DEF3]](s32), [[PTR_ADD]](p0) :: (store 4 + 4)
%5:_(p0) = G_IMPLICIT_DEF
%0:_(s1) = G_IMPLICIT_DEF
G_STORE %0, %5 ::(store 1)

4
unittests/CodeGen/GlobalISel/LegalizerTest.cpp
View File

@ -72,7 +72,7 @@ TEST_F(AArch64GISelMITest, BasicLegalizerTest) {
CHECK-NEXT: [[LOAD_0:%[0-9]+]]:_(s16) = G_LOAD %vptr:_(p0) :: (load 1)
CHECK-NEXT: [[OFFSET_1:%[0-9]+]]:_(s64) = G_CONSTANT i64 1
CHECK-NEXT: [[VPTR_1:%[0-9]+]]:_(p0) = G_PTR_ADD %vptr:_, [[OFFSET_1]]:_(s64)
CHECK-NEXT: [[LOAD_1:%[0-9]+]]:_(s16) = G_LOAD [[VPTR_1]]:_(p0) :: (load 1)
CHECK-NEXT: [[LOAD_1:%[0-9]+]]:_(s16) = G_LOAD [[VPTR_1]]:_(p0) :: (load 1 + 1)
CHECK-NEXT: [[V0:%[0-9]+]]:_(s16) = COPY [[LOAD_0]]:_(s16)
CHECK-NEXT: [[V1:%[0-9]+]]:_(s16) = COPY [[LOAD_1]]:_(s16)
CHECK-NEXT: %v:_(<2 x s8>) = G_BUILD_VECTOR_TRUNC [[V0]]:_(s16), [[V1]]:_(s16)
@ -203,7 +203,7 @@ TEST_F(AArch64GISelMITest, UnorderedArtifactCombiningManyCopiesTest) {
CHECK-NEXT: [[LOAD_0:%[0-9]+]]:_(s16) = G_LOAD %vptr:_(p0) :: (load 1)
CHECK-NEXT: [[OFFSET_1:%[0-9]+]]:_(s64) = G_CONSTANT i64 1
CHECK-NEXT: [[VPTR_1:%[0-9]+]]:_(p0) = G_PTR_ADD %vptr:_, [[OFFSET_1]]:_(s64)
CHECK-NEXT: [[LOAD_1:%[0-9]+]]:_(s16) = G_LOAD [[VPTR_1]]:_(p0) :: (load 1)
CHECK-NEXT: [[LOAD_1:%[0-9]+]]:_(s16) = G_LOAD [[VPTR_1]]:_(p0) :: (load 1 + 1)
CHECK-NEXT: [[FF_MASK:%[0-9]+]]:_(s32) = G_CONSTANT i32 255
CHECK-NEXT: [[V0_EXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LOAD_0]]:_(s16)
CHECK-NEXT: %v0_zext:_(s32) = G_AND [[V0_EXT]]:_, [[FF_MASK]]:_