llvm-project/llvm/test/CodeGen/AMDGPU/vgpr-spill-fi-skip-processing-stack-arg-dbg-value.mir
Christudasan Devadasan a8d7ad70aa [AMDGPU] Skip stack-arg dbg objects while fixing the dead frame indices
Both SGPR->VGPR and VGPR->AGPR spilling code give a fixup to the
spill frame indices referred in debug instructions so that they
can be entirely removed. We should skip the stack argument debug
objects while looking inside the bitvector with FI as the index
that tracks the spill indices being processed. The stack args will
have negative indices and would crash while accessing the bitvector.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D137277
2022-11-04 15:28:35 +05:30

57 lines
2.5 KiB
YAML

# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx908 -amdgpu-spill-vgpr-to-agpr=true -verify-machineinstrs -run-pass=prologepilog -o - %s | FileCheck %s
# After handling the VGPR spill to AGPR copy, we replace the dead frame index in the DBG_VALUE instruction with reg 0.
# Skip looking for frame indices in the debug value instruction for incoming arguments passed via stack. The test would crash otherwise.
# It is safe to skip the fixed stack objects as they will never become the spill objects.
--- |
define amdgpu_kernel void @test() { ret void }
!0 = distinct !DICompileUnit(language: DW_LANG_C99, file: !4, producer: "llvm", isOptimized: true, runtimeVersion: 0, emissionKind: FullDebug, enums: !4, retainedTypes: !4)
!1 = !DILocalVariable(name: "a", scope: !2, file: !4, line: 126, type: !6)
!2 = distinct !DISubprogram(name: "test", scope: !4, file: !4, line: 1, type: !3, isLocal: false, isDefinition: true, scopeLine: 2, flags: DIFlagPrototyped, isOptimized: true, unit: !0, retainedNodes: !5)
!3 = !DISubroutineType(types: !4)
!4 = !{null}
!5 = !{!1}
!6 = !DIDerivedType(tag: DW_TAG_pointer_type, baseType: !7, size: 64, align: 32)
!7 = !DIBasicType(name: "int", size: 32, align: 32, encoding: DW_ATE_signed)
!8 = !DIExpression()
!9 = !DILocation(line: 10, column: 9, scope: !2)
...
---
name: test
tracksRegLiveness: true
frameInfo:
maxAlignment: 4
fixedStack:
- { id: 0, type: default, offset: 4, size: 4, alignment: 4, stack-id: default }
stack:
- { id: 0, type: spill-slot, size: 4, alignment: 4 }
machineFunctionInfo:
maxKernArgAlign: 4
isEntryFunction: true
waveLimiter: true
scratchRSrcReg: '$sgpr96_sgpr97_sgpr98_sgpr99'
stackPtrOffsetReg: '$sgpr32'
frameOffsetReg: '$sgpr33'
hasSpilledVGPRs: true
argumentInfo:
privateSegmentBuffer: { reg: '$sgpr0_sgpr1_sgpr2_sgpr3' }
dispatchPtr: { reg: '$sgpr4_sgpr5' }
kernargSegmentPtr: { reg: '$sgpr6_sgpr7' }
workGroupIDX: { reg: '$sgpr8' }
privateSegmentWaveByteOffset: { reg: '$sgpr9' }
body: |
; CHECK-LABEL: name: test
; CHECK: bb.0:
; CHECK: DBG_VALUE $noreg, 0
bb.0:
$vgpr2 = IMPLICIT_DEF
SI_SPILL_V32_SAVE $vgpr2, %stack.0, $sgpr32, 0, implicit $exec :: (store (s32) into %stack.0, align 4, addrspace 5)
DBG_VALUE %fixed-stack.0, 0, !1, !8, debug-location !9
bb.1:
renamable $vgpr2 = SI_SPILL_V32_RESTORE %stack.0, $sgpr32, 0, implicit $exec :: (load (s32) from %stack.0, align 4, addrspace 5)
S_ENDPGM 0