llvm/test/CodeGen/AMDGPU/fold-fi-mubuf.mir

# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
# RUN: llc -mtriple=amdgcn-amd-amdhsa -verify-machineinstrs -run-pass si-fold-operands,dead-mi-elimination  %s -o - | FileCheck -check-prefix=GCN %s

# Kernels can have no FP
---
name: kernel_no_fold_fi_non_stack_rsrc_and_soffset
tracksRegLiveness: true
frameInfo:
  maxAlignment:    4
  localFrameSize:  4
stack:
  - { id: 0, size: 4, alignment: 4, local-offset: 0 }
machineFunctionInfo:
  isEntryFunction: true
  scratchRSrcReg:  '$sgpr96_sgpr97_sgpr98_sgpr99'
  stackPtrOffsetReg: '$sgpr32'
body:             |
  bb.0:
    liveins: $sgpr12_sgpr13_sgpr14_sgpr15

    ; GCN-LABEL: name: kernel_no_fold_fi_non_stack_rsrc_and_soffset
    ; GCN: liveins: $sgpr12_sgpr13_sgpr14_sgpr15
    ; GCN-NEXT: {{  $}}
    ; GCN-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr12_sgpr13_sgpr14_sgpr15
    ; GCN-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
    ; GCN-NEXT: [[BUFFER_LOAD_DWORD_IDXEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_IDXEN [[V_MOV_B32_e32_]], [[COPY]], 0, 0, 0, 0, implicit $exec
    ; GCN-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_IDXEN]]
    ; GCN-NEXT: SI_RETURN_TO_EPILOG $vgpr0
    %0:sgpr_128 = COPY $sgpr12_sgpr13_sgpr14_sgpr15
    %1:sreg_32_xm0 = S_MOV_B32 0
    %2:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
    %3:vgpr_32 = BUFFER_LOAD_DWORD_IDXEN %2, %0, %1, 0, 0, 0, implicit $exec
    $vgpr0 = COPY %3
    SI_RETURN_TO_EPILOG $vgpr0

...

---
name: kernel_no_fold_fi_non_stack_rsrc
tracksRegLiveness: true
frameInfo:
  maxAlignment:    4
  localFrameSize:  4
stack:
  - { id: 0, size: 4, alignment: 4, local-offset: 0 }
machineFunctionInfo:
  isEntryFunction: true
  scratchRSrcReg:  '$sgpr96_sgpr97_sgpr98_sgpr99'
  stackPtrOffsetReg: '$sgpr32'
body:             |
  bb.0:
    liveins: $sgpr12_sgpr13_sgpr14_sgpr15

    ; GCN-LABEL: name: kernel_no_fold_fi_non_stack_rsrc
    ; GCN: liveins: $sgpr12_sgpr13_sgpr14_sgpr15
    ; GCN-NEXT: {{  $}}
    ; GCN-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr12_sgpr13_sgpr14_sgpr15
    ; GCN-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
    ; GCN-NEXT: [[BUFFER_LOAD_DWORD_IDXEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_IDXEN [[V_MOV_B32_e32_]], [[COPY]], 0, 0, 0, 0, implicit $exec
    ; GCN-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_IDXEN]]
    ; GCN-NEXT: SI_RETURN_TO_EPILOG $vgpr0
    %0:sgpr_128 = COPY $sgpr12_sgpr13_sgpr14_sgpr15
    %2:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
    %3:vgpr_32 = BUFFER_LOAD_DWORD_IDXEN %2, %0, 0, 0, 0, 0, implicit $exec
    $vgpr0 = COPY %3
    SI_RETURN_TO_EPILOG $vgpr0

...

---
name: kernel_no_fold_fi_non_stack_soffset
tracksRegLiveness: true
frameInfo:
  maxAlignment:    4
  localFrameSize:  4
stack:
  - { id: 0, size: 4, alignment: 4, local-offset: 0 }
machineFunctionInfo:
  isEntryFunction: true
  scratchRSrcReg:  '$sgpr0_sgpr1_sgpr2_sgpr3'
  stackPtrOffsetReg: '$sgpr32'
body:             |
  bb.0:

    ; GCN-LABEL: name: kernel_no_fold_fi_non_stack_soffset
    ; GCN: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
    ; GCN-NEXT: [[V_MOV_B32_e32_1:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 7, implicit $exec
    ; GCN-NEXT: BUFFER_STORE_DWORD_OFFEN [[V_MOV_B32_e32_1]], [[V_MOV_B32_e32_]], $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    ; GCN-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_MOV_B32_e32_]], $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    ; GCN-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
    ; GCN-NEXT: S_ENDPGM 0, implicit $vgpr0
    %0:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
    %1:vgpr_32 = V_MOV_B32_e32 7, implicit $exec
    %2:sreg_32_xm0 = S_MOV_B32 0

    BUFFER_STORE_DWORD_OFFEN %1:vgpr_32, %0:vgpr_32, $sgpr0_sgpr1_sgpr2_sgpr3, %2, 0, 0, 0, implicit $exec
    %3:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN %0:vgpr_32, $sgpr0_sgpr1_sgpr2_sgpr3, %2, 0, 0, 0, implicit $exec
    $vgpr0 = COPY %3
    S_ENDPGM 0, implicit $vgpr0

...

---
name: kernel_fold_fi_mubuf
tracksRegLiveness: true
frameInfo:
  maxAlignment:    4
  localFrameSize:  4
stack:
  - { id: 0, size: 4, alignment: 4, local-offset: 0 }
machineFunctionInfo:
  isEntryFunction: true
  scratchRSrcReg:  '$sgpr0_sgpr1_sgpr2_sgpr3'
  stackPtrOffsetReg: '$sgpr32'
body:             |
  bb.0:

    ; GCN-LABEL: name: kernel_fold_fi_mubuf
    ; GCN: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 7, implicit $exec
    ; GCN-NEXT: BUFFER_STORE_DWORD_OFFEN [[V_MOV_B32_e32_]], %stack.0, $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    ; GCN-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN %stack.0, $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    ; GCN-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
    ; GCN-NEXT: S_ENDPGM 0, implicit $vgpr0
    %0:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
    %1:vgpr_32 = V_MOV_B32_e32 7, implicit $exec

    BUFFER_STORE_DWORD_OFFEN %1:vgpr_32, %0:vgpr_32, $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    %2:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN %0:vgpr_32, $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    $vgpr0 = COPY %2
    S_ENDPGM 0, implicit $vgpr0

...


# Functions have an unswizzled SP/FP relative to the wave offset
---
name: function_no_fold_fi_non_stack_rsrc_and_soffset
tracksRegLiveness: true
frameInfo:
  maxAlignment:    4
  localFrameSize:  4
stack:
  - { id: 0, size: 4, alignment: 4, local-offset: 0 }
machineFunctionInfo:
  isEntryFunction: false
  scratchRSrcReg:  '$sgpr96_sgpr97_sgpr98_sgpr99'
  frameOffsetReg:  '$sgpr32'
  stackPtrOffsetReg: '$sgpr32'
body:             |
  bb.0:
    liveins: $sgpr12_sgpr13_sgpr14_sgpr15

    ; GCN-LABEL: name: function_no_fold_fi_non_stack_rsrc_and_soffset
    ; GCN: liveins: $sgpr12_sgpr13_sgpr14_sgpr15
    ; GCN-NEXT: {{  $}}
    ; GCN-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr12_sgpr13_sgpr14_sgpr15
    ; GCN-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
    ; GCN-NEXT: [[BUFFER_LOAD_DWORD_IDXEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_IDXEN [[V_MOV_B32_e32_]], [[COPY]], 0, 0, 0, 0, implicit $exec
    ; GCN-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_IDXEN]]
    ; GCN-NEXT: SI_RETURN_TO_EPILOG $vgpr0
    %0:sgpr_128 = COPY $sgpr12_sgpr13_sgpr14_sgpr15
    %1:sreg_32_xm0 = S_MOV_B32 0
    %2:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
    %3:vgpr_32 = BUFFER_LOAD_DWORD_IDXEN %2, %0, %1, 0, 0, 0, implicit $exec
    $vgpr0 = COPY %3
    SI_RETURN_TO_EPILOG $vgpr0

...

---
name: function_no_fold_fi_non_stack_rsrc
tracksRegLiveness: true
frameInfo:
  maxAlignment:    4
  localFrameSize:  4
stack:
  - { id: 0, size: 4, alignment: 4, local-offset: 0 }
machineFunctionInfo:
  isEntryFunction: false
  scratchRSrcReg:  '$sgpr96_sgpr97_sgpr98_sgpr99'
  frameOffsetReg:  '$sgpr32'
  stackPtrOffsetReg: '$sgpr32'
body:             |
  bb.0:
    liveins: $sgpr12_sgpr13_sgpr14_sgpr15

    ; GCN-LABEL: name: function_no_fold_fi_non_stack_rsrc
    ; GCN: liveins: $sgpr12_sgpr13_sgpr14_sgpr15
    ; GCN-NEXT: {{  $}}
    ; GCN-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr12_sgpr13_sgpr14_sgpr15
    ; GCN-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
    ; GCN-NEXT: [[BUFFER_LOAD_DWORD_IDXEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_IDXEN [[V_MOV_B32_e32_]], [[COPY]], 0, 0, 0, 0, implicit $exec
    ; GCN-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_IDXEN]]
    ; GCN-NEXT: SI_RETURN_TO_EPILOG $vgpr0
    %0:sgpr_128 = COPY $sgpr12_sgpr13_sgpr14_sgpr15
    %2:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
    %3:vgpr_32 = BUFFER_LOAD_DWORD_IDXEN %2, %0, 0, 0, 0, 0, implicit $exec
    $vgpr0 = COPY %3
    SI_RETURN_TO_EPILOG $vgpr0

...

---
name: function_no_fold_fi_non_stack_soffset
tracksRegLiveness: true
frameInfo:
  maxAlignment:    4
  localFrameSize:  4
stack:
  - { id: 0, size: 4, alignment: 4, local-offset: 0 }
machineFunctionInfo:
  isEntryFunction: false
  scratchRSrcReg:  '$sgpr0_sgpr1_sgpr2_sgpr3'
  frameOffsetReg:  '$sgpr32'
  stackPtrOffsetReg: '$sgpr32'
body:             |
  bb.0:

    ; GCN-LABEL: name: function_no_fold_fi_non_stack_soffset
    ; GCN: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 7, implicit $exec
    ; GCN-NEXT: BUFFER_STORE_DWORD_OFFEN [[V_MOV_B32_e32_]], %stack.0, $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    ; GCN-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN %stack.0, $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    ; GCN-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
    ; GCN-NEXT: S_ENDPGM 0, implicit $vgpr0
    %0:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
    %1:vgpr_32 = V_MOV_B32_e32 7, implicit $exec

    BUFFER_STORE_DWORD_OFFEN %1:vgpr_32, %0:vgpr_32, $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    %2:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN %0:vgpr_32, $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    $vgpr0 = COPY %2
    S_ENDPGM 0, implicit $vgpr0

...

---
name: function_fold_fi_mubuf_wave_relative
tracksRegLiveness: true
frameInfo:
  maxAlignment:    4
  localFrameSize:  4
stack:
  - { id: 0, size: 4, alignment: 4, local-offset: 0 }
machineFunctionInfo:
  isEntryFunction: false
  scratchRSrcReg:  '$sgpr0_sgpr1_sgpr2_sgpr3'
  frameOffsetReg:  '$sgpr32'
  stackPtrOffsetReg: '$sgpr32'
body:             |
  bb.0:

    ; GCN-LABEL: name: function_fold_fi_mubuf_wave_relative
    ; GCN: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 7, implicit $exec
    ; GCN-NEXT: BUFFER_STORE_DWORD_OFFEN [[V_MOV_B32_e32_]], %stack.0, $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    ; GCN-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN %stack.0, $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    ; GCN-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
    ; GCN-NEXT: S_ENDPGM 0, implicit $vgpr0
    %0:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
    %1:vgpr_32 = V_MOV_B32_e32 7, implicit $exec

    BUFFER_STORE_DWORD_OFFEN %1:vgpr_32, %0:vgpr_32, $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    %2:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN %0:vgpr_32, $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    $vgpr0 = COPY %2
    S_ENDPGM 0, implicit $vgpr0

...

---
name: function_fold_fi_mubuf_stack_relative
tracksRegLiveness: true
frameInfo:
  maxAlignment:    4
  localFrameSize:  4
stack:
  - { id: 0, size: 4, alignment: 4, local-offset: 0 }
machineFunctionInfo:
  isEntryFunction: false
  scratchRSrcReg:  '$sgpr0_sgpr1_sgpr2_sgpr3'
  frameOffsetReg:  '$sgpr32'
  stackPtrOffsetReg: '$sgpr32'
body:             |
  bb.0:

    ; GCN-LABEL: name: function_fold_fi_mubuf_stack_relative
    ; GCN: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 7, implicit $exec
    ; GCN-NEXT: BUFFER_STORE_DWORD_OFFEN [[V_MOV_B32_e32_]], %stack.0, $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    ; GCN-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN %stack.0, $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    ; GCN-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
    ; GCN-NEXT: S_ENDPGM 0, implicit $vgpr0
    %0:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
    %1:vgpr_32 = V_MOV_B32_e32 7, implicit $exec

    BUFFER_STORE_DWORD_OFFEN %1:vgpr_32, %0:vgpr_32, $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    %2:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN %0:vgpr_32, $sgpr0_sgpr1_sgpr2_sgpr3, 0, 0, 0, 0, implicit $exec
    $vgpr0 = COPY %2
    S_ENDPGM 0, implicit $vgpr0

...