3
3
# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx908 %s -o - -run-pass prologepilog,machine-cp -verify-machineinstrs | FileCheck -check-prefix=GFX908-PEI-MACHINECP %s
4
4
5
5
# When VGPRs are available for spilling, prologepilog marks the tuple implicit-def as well as implicit in the first spill instruction.
6
- # As a consequence, machine-cp would NOT delete agpr2 copy here.
7
6
8
7
---
9
8
name : agpr-spill-to-vgpr-machine-cp
10
9
tracksRegLiveness : true
11
10
stack :
12
11
- { id: 0, name: '', type: spill-slot, offset: 0, size: 128, alignment: 4 }
13
12
machineFunctionInfo :
13
+ isEntryFunction : true
14
14
scratchRSrcReg : $sgpr0_sgpr1_sgpr2_sgpr3
15
15
stackPtrOffsetReg : ' $sgpr32'
16
16
hasSpilledVGPRs : true
@@ -43,15 +43,16 @@ body: |
43
43
S_ENDPGM 0
44
44
...
45
45
46
- # When VGPRs are NOT available for spilling (stack is used), prologepilog marks the tuple implicit-def only and NOT implicit.
47
- # As a consequence, machine-cp would delete agpr2 copy here. Presently, this is incorrect behavior .
46
+ # When VGPRs are NOT available for spilling (stack is used), prologepilog should also mark the tuple implicit-def and implicit (similar to above usecase) .
47
+ # As a consequence, machine-cp would not delete agpr2 copy here.
48
48
49
49
---
50
50
name : agpr-spill-to-vgpr-to-stack-machine-cp
51
51
tracksRegLiveness : true
52
52
stack :
53
53
- { id: 0, name: '', type: spill-slot, offset: 0, size: 128, alignment: 4 }
54
54
machineFunctionInfo :
55
+ isEntryFunction : true
55
56
scratchRSrcReg : $sgpr0_sgpr1_sgpr2_sgpr3
56
57
stackPtrOffsetReg : ' $sgpr32'
57
58
hasSpilledVGPRs : true
@@ -60,29 +61,34 @@ body: |
60
61
successors:
61
62
liveins: $vgpr0, $vgpr1
62
63
; GFX908-PEI-LABEL: name: agpr-spill-to-vgpr-to-stack-machine-cp
63
- ; GFX908-PEI: liveins: $vgpr0, $vgpr1, $vgpr18, $vgpr19, $vgpr20, $vgpr21, $vgpr22, $vgpr23, $vgpr24, $vgpr25, $vgpr26, $vgpr27, $vgpr28, $vgpr29, $vgpr30, $vgpr31, $vgpr32, $vgpr33, $vgpr34, $vgpr35, $vgpr36, $vgpr37, $vgpr38, $vgpr39, $vgpr48, $vgpr49, $vgpr50, $vgpr51, $vgpr52, $vgpr53, $vgpr54, $vgpr55
64
+ ; GFX908-PEI: liveins: $vgpr0, $vgpr1, $vgpr18, $vgpr19, $vgpr20, $vgpr21, $vgpr22, $vgpr23, $vgpr24, $vgpr25, $vgpr26, $vgpr27, $vgpr28, $vgpr29, $vgpr30, $vgpr31, $vgpr32, $vgpr33, $vgpr34, $vgpr35, $vgpr36, $vgpr37, $vgpr38, $vgpr39, $vgpr48, $vgpr49, $vgpr50, $vgpr51, $vgpr52, $vgpr53, $vgpr54, $vgpr55, $sgpr0_sgpr1_sgpr2_sgpr3
64
65
; GFX908-PEI-NEXT: {{ $}}
66
+ ; GFX908-PEI-NEXT: $sgpr0 = S_ADD_U32 $sgpr0, $noreg, implicit-def $scc, implicit-def $sgpr0_sgpr1_sgpr2_sgpr3
67
+ ; GFX908-PEI-NEXT: $sgpr1 = S_ADDC_U32 $sgpr1, 0, implicit-def dead $scc, implicit $scc, implicit-def $sgpr0_sgpr1_sgpr2_sgpr3
65
68
; GFX908-PEI-NEXT: renamable $agpr0 = COPY renamable $vgpr0, implicit $exec
66
69
; GFX908-PEI-NEXT: renamable $agpr2 = COPY renamable $vgpr1, implicit $exec
67
70
; GFX908-PEI-NEXT: $vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9 = IMPLICIT_DEF
68
71
; GFX908-PEI-NEXT: $vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17 = IMPLICIT_DEF
69
- ; GFX908-PEI-NEXT: $vgpr40 = V_ACCVGPR_READ_B32_e64 $agpr0, implicit $exec, implicit-def $agpr0_agpr1_agpr2
70
- ; GFX908-PEI-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr40, $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32 , 0, 0, 0, implicit $exec, implicit $agpr0_agpr1_agpr2 :: (store (s32) into %stack.0, addrspace 5)
72
+ ; GFX908-PEI-NEXT: $vgpr40 = V_ACCVGPR_READ_B32_e64 $agpr0, implicit $exec, implicit-def $agpr0_agpr1_agpr2, implicit $agpr0_agpr1_agpr2
73
+ ; GFX908-PEI-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr40, $sgpr0_sgpr1_sgpr2_sgpr3, 0 , 0, 0, 0, implicit $exec, implicit $agpr0_agpr1_agpr2 :: (store (s32) into %stack.0, addrspace 5)
71
74
; GFX908-PEI-NEXT: $vgpr40 = V_ACCVGPR_READ_B32_e64 $agpr1, implicit $exec
72
- ; GFX908-PEI-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr40, $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32 , 4, 0, 0, implicit $exec :: (store (s32) into %stack.0 + 4, addrspace 5)
75
+ ; GFX908-PEI-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr40, $sgpr0_sgpr1_sgpr2_sgpr3, 0 , 4, 0, 0, implicit $exec :: (store (s32) into %stack.0 + 4, addrspace 5)
73
76
; GFX908-PEI-NEXT: $vgpr55 = V_ACCVGPR_READ_B32_e64 $agpr2, implicit $exec, implicit $agpr0_agpr1_agpr2
74
77
; GFX908-PEI-NEXT: S_ENDPGM 0
75
78
;
76
79
; GFX908-PEI-MACHINECP-LABEL: name: agpr-spill-to-vgpr-to-stack-machine-cp
77
- ; GFX908-PEI-MACHINECP: liveins: $vgpr0, $vgpr1, $vgpr18, $vgpr19, $vgpr20, $vgpr21, $vgpr22, $vgpr23, $vgpr24, $vgpr25, $vgpr26, $vgpr27, $vgpr28, $vgpr29, $vgpr30, $vgpr31, $vgpr32, $vgpr33, $vgpr34, $vgpr35, $vgpr36, $vgpr37, $vgpr38, $vgpr39, $vgpr48, $vgpr49, $vgpr50, $vgpr51, $vgpr52, $vgpr53, $vgpr54, $vgpr55
80
+ ; GFX908-PEI-MACHINECP: liveins: $vgpr0, $vgpr1, $vgpr18, $vgpr19, $vgpr20, $vgpr21, $vgpr22, $vgpr23, $vgpr24, $vgpr25, $vgpr26, $vgpr27, $vgpr28, $vgpr29, $vgpr30, $vgpr31, $vgpr32, $vgpr33, $vgpr34, $vgpr35, $vgpr36, $vgpr37, $vgpr38, $vgpr39, $vgpr48, $vgpr49, $vgpr50, $vgpr51, $vgpr52, $vgpr53, $vgpr54, $vgpr55, $sgpr0_sgpr1_sgpr2_sgpr3
78
81
; GFX908-PEI-MACHINECP-NEXT: {{ $}}
82
+ ; GFX908-PEI-MACHINECP-NEXT: $sgpr0 = S_ADD_U32 $sgpr0, $noreg, implicit-def $scc, implicit-def $sgpr0_sgpr1_sgpr2_sgpr3
83
+ ; GFX908-PEI-MACHINECP-NEXT: $sgpr1 = S_ADDC_U32 $sgpr1, 0, implicit-def dead $scc, implicit $scc, implicit-def $sgpr0_sgpr1_sgpr2_sgpr3
79
84
; GFX908-PEI-MACHINECP-NEXT: renamable $agpr0 = COPY renamable $vgpr0, implicit $exec
85
+ ; GFX908-PEI-MACHINECP-NEXT: renamable $agpr2 = COPY renamable $vgpr1, implicit $exec
80
86
; GFX908-PEI-MACHINECP-NEXT: $vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9 = IMPLICIT_DEF
81
87
; GFX908-PEI-MACHINECP-NEXT: $vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17 = IMPLICIT_DEF
82
- ; GFX908-PEI-MACHINECP-NEXT: $vgpr40 = V_ACCVGPR_READ_B32_e64 $agpr0, implicit $exec, implicit-def $agpr0_agpr1_agpr2
83
- ; GFX908-PEI-MACHINECP-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr40, $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32 , 0, 0, 0, implicit $exec, implicit $agpr0_agpr1_agpr2 :: (store (s32) into %stack.0, addrspace 5)
88
+ ; GFX908-PEI-MACHINECP-NEXT: $vgpr40 = V_ACCVGPR_READ_B32_e64 $agpr0, implicit $exec, implicit-def $agpr0_agpr1_agpr2, implicit $agpr0_agpr1_agpr2
89
+ ; GFX908-PEI-MACHINECP-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr40, $sgpr0_sgpr1_sgpr2_sgpr3, 0 , 0, 0, 0, implicit $exec, implicit $agpr0_agpr1_agpr2 :: (store (s32) into %stack.0, addrspace 5)
84
90
; GFX908-PEI-MACHINECP-NEXT: $vgpr40 = V_ACCVGPR_READ_B32_e64 $agpr1, implicit $exec
85
- ; GFX908-PEI-MACHINECP-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr40, $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32 , 4, 0, 0, implicit $exec :: (store (s32) into %stack.0 + 4, addrspace 5)
91
+ ; GFX908-PEI-MACHINECP-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr40, $sgpr0_sgpr1_sgpr2_sgpr3, 0 , 4, 0, 0, implicit $exec :: (store (s32) into %stack.0 + 4, addrspace 5)
86
92
; GFX908-PEI-MACHINECP-NEXT: $vgpr55 = V_ACCVGPR_READ_B32_e64 $agpr2, implicit $exec, implicit $agpr0_agpr1_agpr2
87
93
; GFX908-PEI-MACHINECP-NEXT: S_ENDPGM 0
88
94
renamable $agpr0 = COPY renamable $vgpr0, implicit $exec
0 commit comments